Skip to content

Beautiful, AI-native markdown IDE and LLM wiki

One checkbox to connect your AI tools (#3538)

mike-inkeep375c56670812
diff --git a/.changeset/conditional-fish-path-install.md b/.changeset/conditional-fish-path-install.md
new file mode 100644
index 000000000..2e794ba42
--- /dev/null
+++ b/.changeset/conditional-fish-path-install.md
@@ -0,0 +1,5 @@
+---
+'@inkeep/open-knowledge': patch
+---
+
+For fresh PATH setups, the desktop app now adds the `ok` command to Fish only when Fish is the login shell or has independent usage or configuration evidence. New Zsh and Bash setups no longer create an unused Fish configuration file; files recorded by older installs remain available to Settings and `ok uninstall` for cleanup.
diff --git a/.changeset/one-checkbox-ai-tool-setup.md b/.changeset/one-checkbox-ai-tool-setup.md
new file mode 100644
index 000000000..19d960ba3
--- /dev/null
+++ b/.changeset/one-checkbox-ai-tool-setup.md
@@ -0,0 +1,30 @@
+---
+'@inkeep/open-knowledge': patch
+---
+
+Setting up your AI tools is now one checkbox instead of a checklist.
+
+The first-launch screen used to ask for a row-by-row audit — a checkbox per tool
+(including tools you don't have), a checkbox per skill, and a Skip button that
+meant something subtly different from unchecking everything. Now it's a single
+pre-checked box covering every tool we found, with a "What this changes" expander
+listing the exact files it writes. Adding `ok` to your terminal stays a separate
+checkbox, since that's an unrelated decision. The same treatment applies when you
+create a project: the AI-tool choice moves out of Advanced settings and becomes
+one checkbox with the same disclosure.
+
+You see more than before, not less. The checkbox names every tool it will write
+to whether or not you expand it, and if setup would replace an existing
+OpenKnowledge entry, that warning sits right next to the checkbox.
+
+It also only names tools it can actually set up. Creating a project used to
+offer tools that take no per-project setup at all — Claude Desktop, OpenClaw,
+Antigravity, LM Studio, Hermes — where ticking the box did nothing. Copilot is
+listed once its OpenKnowledge connection exists, since until then its project
+skill can't load.
+
+Two behavior changes worth knowing: onboarding no longer installs the
+`open-knowledge-write-skill` bundle — it's an authoring convenience, so it now
+waits for you to turn it on in Settings, and any copy you already have is left
+alone. And declining setup never removes anything: it writes nothing and tears
+nothing down. Removal lives in Settings, where the row tells you what it removes.
diff --git a/packages/app/src/components/CreateProjectDialog.runtime.dom.test.tsx b/packages/app/src/components/CreateProjectDialog.runtime.dom.test.tsx
index ab504524d..26708c00f 100644
--- a/packages/app/src/components/CreateProjectDialog.runtime.dom.test.tsx
+++ b/packages/app/src/components/CreateProjectDialog.runtime.dom.test.tsx
@@ -1,4 +1,4 @@
-import { ALL_EDITOR_IDS, EDITOR_LABELS } from '@inkeep/open-knowledge-core';
+import { EDITOR_LABELS } from '@inkeep/open-knowledge-core';
import { cleanup, fireEvent, render, screen, waitFor } from '@testing-library/react';
import userEvent from '@testing-library/user-event';
import { afterEach, beforeEach, describe, expect, test, vi } from 'vitest';
@@ -77,6 +77,15 @@ const UNDETECTED: OkMcpWiringEditorId = 'codex';
function makeBridge() {
let pickedParent: string | null = PARENT;
let detectedEditorIdsImpl = (): Promise<OkMcpWiringEditorId[]> => Promise.resolve([...DETECTED]);
+ /** Per-editor user-global MCP state. Empty = no editor has an OK entry yet. */
+ let editorStatesImpl = (): Array<{
+ id: OkMcpWiringEditorId;
+ label: string;
+ detected: boolean;
+ state: 'installed' | 'not-installed' | 'foreign' | 'unmanageable';
+ configPath: string | null;
+ entryLocator: string;
+ }> => [];
let defaultRootImpl = (): Promise<string> => Promise.resolve(PARENT);
let folderStateImpl = async (_path: string): Promise<OkFolderState> => 'free';
let createNewImpl = (): Promise<void> => Promise.resolve();
@@ -109,11 +118,13 @@ function makeBridge() {
}),
},
integrations: {
- // The dialog reads only `detectedEditorIds` off this snapshot; the rest of
- // the shape exists so the fake matches the real status contract.
+ // The dialog reads `detectedEditorIds` plus each editor's `state` (which
+ // tells it whether a user-global OpenKnowledge entry exists — Copilot's
+ // project skill is gated on that). The rest of the shape exists so the
+ // fake matches the real status contract.
status: vi.fn(async () => ({
available: true,
- editors: [],
+ editors: editorStatesImpl(),
path: { shellDetected: false, rcFilesToTouch: [], installed: false },
skills: [],
detectedEditorIds: await detectedEditorIdsImpl(),
@@ -153,6 +164,9 @@ function makeBridge() {
setDetectedEditorsImpl: (next: () => Promise<OkMcpWiringEditorId[]>) => {
detectedEditorIdsImpl = next;
},
+ setEditorStatesImpl: (next: typeof editorStatesImpl) => {
+ editorStatesImpl = next;
+ },
setDefaultProjectsRootImpl: (next: () => Promise<string>) => {
defaultRootImpl = next;
},
@@ -208,7 +222,7 @@ describe('CreateProjectDialog runtime wiring', () => {
cleanup();
});
- test('detected agent harnesses start selected, undetected ones do not, and edits ride along', async () => {
+ test('the AI-tools row is always visible, names the detected tools, and rides along on submit', async () => {
const stub = await renderDialog();
const form = screen.getByTestId('create-project-form') as HTMLFormElement;
@@ -235,36 +249,33 @@ describe('CreateProjectDialog runtime wiring', () => {
await waitForLocationHydrate();
- // Config sharing and the editor controls both live inside the collapsed
- // "Advanced settings" section (Radix unmounts collapsed content), so
- // neither is in the DOM until the section is expanded.
+ // Config sharing still lives inside the collapsed "Advanced settings"
+ // section (Radix unmounts collapsed content), so it is not in the DOM until
+ // expanded. The AI-tools decision no longer does — it decides whether the
+ // project is reachable from the user's agents at all.
expect(screen.queryByTestId('create-sharing')).toBeNull();
- expect(screen.queryByTestId('create-editor-cursor')).toBeNull();
- fireEvent.click(screen.getByTestId('create-advanced-trigger'));
- expect(screen.getByTestId('create-sharing')).not.toBeNull();
-
- // Every editor still renders a labelled row; only the DETECTED ones come
- // pre-checked. A tool whose host root does not exist is never pre-selected,
- // so we never create a config dir for something the user doesn't have.
await waitFor(() => {
- expect(screen.getByTestId('create-editor-claude').getAttribute('aria-checked')).toBe('true');
+ expect(screen.getByTestId('create-editors-checkbox').getAttribute('aria-checked')).toBe(
+ 'true',
+ );
});
- for (const id of ALL_EDITOR_IDS) {
- const checkbox = screen.getByTestId(`create-editor-${id}`);
- expect(checkbox.closest('label')?.textContent).toContain(EDITOR_LABELS[id]);
- expect(checkbox.getAttribute('aria-checked')).toBe(DETECTED.includes(id) ? 'true' : 'false');
- }
- expect(screen.getByTestId(`create-editor-${UNDETECTED}`).getAttribute('aria-checked')).toBe(
- 'false',
- );
+ fireEvent.click(screen.getByTestId('create-advanced-trigger'));
+ expect(screen.getByTestId('create-sharing')).not.toBeNull();
- // Ticking an undetected tool adds it; unticking a detected one drops it.
- fireEvent.click(screen.getByTestId(`create-editor-${UNDETECTED}`));
- expect(screen.getByTestId(`create-editor-${UNDETECTED}`).getAttribute('aria-checked')).toBe(
- 'true',
- );
- fireEvent.click(screen.getByTestId('create-editor-cursor'));
- expect(screen.getByTestId('create-editor-cursor').getAttribute('aria-checked')).toBe('false');
+ // Consent integrity: collapsed, the label already names every tool that
+ // gets written to, and names no tool that doesn't.
+ const summary = screen.getByTestId('create-editors-summary').textContent ?? '';
+ for (const id of DETECTED) expect(summary).toContain(EDITOR_LABELS[id]);
+ expect(summary).not.toContain(EDITOR_LABELS[UNDETECTED]);
+ // Undetected tools have no row anywhere — nothing is written for them.
+ expect(screen.queryByTestId(`create-editor-${UNDETECTED}`)).toBeNull();
+
+ // The disclosure names the exact project-relative artifacts.
+ fireEvent.click(screen.getByTestId('create-editors-details-toggle'));
+ const details = screen.getByTestId('create-editors-details');
+ expect(details.textContent).toContain('.mcp.json');
+ expect(details.textContent).toContain('.claude/skills/open-knowledge/');
+ expect(details.textContent).toContain('.cursor/mcp.json');
fireEvent.click(cancel);
expect(stub.onOpenChange).toHaveBeenCalledWith(false);
@@ -282,21 +293,36 @@ describe('CreateProjectDialog runtime wiring', () => {
expect(submitted?.parent).toBe(PARENT);
expect(submitted?.name).toBe(PROJECT_NAME);
expect(submitted?.sharing).toBe('shared');
- expect([...(submitted?.editors ?? [])].sort()).toEqual(['claude', UNDETECTED].sort());
+ expect([...(submitted?.editors ?? [])].sort()).toEqual([...DETECTED].sort());
expect(stub.onOpenChange).toHaveBeenLastCalledWith(false);
});
- test('submitting without ever opening Advanced settings wires the detected editors', async () => {
- // The regression this pins: the editor checkboxes live behind a collapsed
- // "Advanced settings" section, so the overwhelmingly common path is
- // name → Create. That path must still produce MCP config + the project
- // skill for the tools the user actually has — an empty `editors` array
- // silently yields a project wired to nothing, and nothing back-fills it.
+ test('the straight-through path wires the detected editors', async () => {
+ // The overwhelmingly common path is name → Create. It must produce MCP
+ // config + the project skill for the tools the user actually has — an empty
+ // `editors` array silently yields a project wired to nothing, and nothing
+ // back-fills it.
+ const stub = await renderDialog();
+ await waitForLocationHydrate();
+
+ await typeProjectName(PROJECT_NAME);
+ await waitForSubmitEnabled();
+ fireEvent.click(screen.getByTestId('create-submit'));
+
+ await waitFor(() => {
+ expect(stub.createNewCalls).toHaveLength(1);
+ });
+ expect([...(stub.createNewCalls[0]?.editors ?? [])].sort()).toEqual([...DETECTED].sort());
+ });
+
+ test('unchecking the row creates the project without wiring anything', async () => {
const stub = await renderDialog();
await waitForLocationHydrate();
- // Advanced is never expanded in this test.
- expect(screen.queryByTestId('create-editor-claude')).toBeNull();
+ await waitFor(() => {
+ expect(screen.queryByTestId('create-editors-checkbox')).not.toBeNull();
+ });
+ fireEvent.click(screen.getByTestId('create-editors-checkbox'));
await typeProjectName(PROJECT_NAME);
await waitForSubmitEnabled();
@@ -305,17 +331,168 @@ describe('CreateProjectDialog runtime wiring', () => {
await waitFor(() => {
expect(stub.createNewCalls).toHaveLength(1);
});
+ expect(stub.createNewCalls[0]?.editors).toEqual([]);
+ });
+
+ test('a detected user-global-only tool is neither named nor submitted', async () => {
+ // Claude Desktop has no project MCP config and no project skill root, so
+ // every project writer returns `skipped-unsupported` for it. Detection
+ // still finds it (its host root exists), which is exactly why the filter
+ // has to be on what gets WRITTEN rather than on what was detected.
+ const stub = makeBridge();
+ stub.setDetectedEditorsImpl(() => Promise.resolve(['claude', 'claude-desktop']));
+ await renderDialog(stub);
+ await waitForLocationHydrate();
+
+ await waitFor(() => {
+ expect(screen.queryByTestId('create-editors-checkbox')).not.toBeNull();
+ });
+ const summary = screen.getByTestId('create-editors-summary').textContent ?? '';
+ expect(summary).toContain(EDITOR_LABELS.claude);
+ expect(summary).not.toContain(EDITOR_LABELS['claude-desktop']);
+
+ await typeProjectName(PROJECT_NAME);
+ await waitForSubmitEnabled();
+ fireEvent.click(screen.getByTestId('create-submit'));
+ await waitFor(() => {
+ expect(stub.createNewCalls).toHaveLength(1);
+ });
+ expect(stub.createNewCalls[0]?.editors).toEqual(['claude']);
+ });
+
+ test('Copilot is dropped until its user-global entry exists, then included', async () => {
+ // Copilot's project skill (`.github/skills`) is refused by
+ // `isProjectSkillPrerequisiteMet` until Copilot's USER-GLOBAL OpenKnowledge
+ // entry is present, and it has no project MCP config — so before that, a
+ // create writes nothing for it and must not say otherwise.
+ const withoutEntry = makeBridge();
+ withoutEntry.setDetectedEditorsImpl(() => Promise.resolve(['claude', 'copilot']));
+ await renderDialog(withoutEntry);
+ await waitForLocationHydrate();
+ await waitFor(() => {
+ expect(screen.queryByTestId('create-editors-checkbox')).not.toBeNull();
+ });
+ expect(screen.getByTestId('create-editors-summary').textContent ?? '').not.toContain(
+ EDITOR_LABELS.copilot,
+ );
+
+ cleanup();
+
+ const withEntry = makeBridge();
+ withEntry.setDetectedEditorsImpl(() => Promise.resolve(['claude', 'copilot']));
+ withEntry.setEditorStatesImpl(() => [
+ {
+ id: 'copilot',
+ label: EDITOR_LABELS.copilot,
+ detected: true,
+ state: 'installed',
+ configPath: '~/.copilot/mcp-config.json',
+ entryLocator: 'mcpServers.open-knowledge',
+ },
+ ]);
+ await renderDialog(withEntry);
+ await waitForLocationHydrate();
+ await waitFor(() => {
+ expect(screen.getByTestId('create-editors-summary').textContent ?? '').toContain(
+ EDITOR_LABELS.copilot,
+ );
+ });
+
+ await typeProjectName(PROJECT_NAME);
+ await waitForSubmitEnabled();
+ fireEvent.click(screen.getByTestId('create-submit'));
+ await waitFor(() => {
+ expect(withEntry.createNewCalls).toHaveLength(1);
+ });
+ expect([...(withEntry.createNewCalls[0]?.editors ?? [])].sort()).toEqual(['claude', 'copilot']);
+ });
+
+ test('a foreign Copilot entry is treated as not-connected', async () => {
+ // The renderer reads `state === 'installed'` — deliberately stricter than the
+ // write path, which passes on ANY entry under OpenKnowledge's server name.
+ // A foreign entry means OK's MCP isn't registered, so the skill would point
+ // the agent at tools that aren't there. Widening this filter to
+ // `!== 'not-installed'` is the regression this pins.
+ const stub = makeBridge();
+ stub.setDetectedEditorsImpl(() => Promise.resolve(['claude', 'copilot']));
+ stub.setEditorStatesImpl(() => [
+ {
+ id: 'copilot',
+ label: EDITOR_LABELS.copilot,
+ detected: true,
+ state: 'foreign',
+ configPath: '~/.copilot/mcp-config.json',
+ entryLocator: 'mcpServers.open-knowledge',
+ },
+ ]);
+ await renderDialog(stub);
+ await waitForLocationHydrate();
+
+ await waitFor(() => {
+ expect(screen.queryByTestId('create-editors-checkbox')).not.toBeNull();
+ });
+ expect(screen.getByTestId('create-editors-summary').textContent ?? '').not.toContain(
+ EDITOR_LABELS.copilot,
+ );
+
+ await typeProjectName(PROJECT_NAME);
+ await waitForSubmitEnabled();
+ fireEvent.click(screen.getByTestId('create-submit'));
+ await waitFor(() => {
+ expect(stub.createNewCalls).toHaveLength(1);
+ });
+ expect(stub.createNewCalls[0]?.editors).toEqual(['claude']);
+ });
+
+ test('Create stays blocked while detection is in flight, so no project is wired to nothing', async () => {
+ // The detection probe settles independently of the location/cascade probes,
+ // so Create could otherwise unlock while `detectedEditors` is still null —
+ // and submitting then sends `editors: []`, silently creating a project wired
+ // to nothing while the row still reads "Checking which AI tools you have".
+ let releaseDetection = (): void => {};
+ const stub = makeBridge();
+ stub.setDetectedEditorsImpl(
+ () =>
+ new Promise<OkMcpWiringEditorId[]>((resolve) => {
+ releaseDetection = () => resolve([...DETECTED]);
+ }),
+ );
+ await renderDialog(stub);
+ await waitForLocationHydrate();
+ await typeProjectName(PROJECT_NAME);
+
+ // Location probe has settled; detection has not. Create must stay disabled.
+ await waitFor(() => {
+ expect(screen.getByTestId('create-editors-status').getAttribute('data-status')).toBe(
+ 'probing',
+ );
+ });
+ expect((screen.getByTestId('create-submit') as HTMLButtonElement).disabled).toBe(true);
+
+ releaseDetection();
+ await waitForSubmitEnabled();
+ fireEvent.click(screen.getByTestId('create-submit'));
+ await waitFor(() => {
+ expect(stub.createNewCalls).toHaveLength(1);
+ });
+ // The write set is the real one, not the empty list the race would produce.
expect([...(stub.createNewCalls[0]?.editors ?? [])].sort()).toEqual([...DETECTED].sort());
});
- test('a failed detection probe leaves the selection empty rather than guessing', async () => {
+ test('a failed detection probe settles empty rather than guessing', async () => {
// Degrade toward writing nothing, never toward creating host roots for
- // tools we could not confirm. Advanced settings remains the manual escape.
+ // tools we could not confirm — and say so, rather than hanging on the
+ // in-flight placeholder forever.
const stub = makeBridge();
stub.setDetectedEditorsImpl(() => Promise.reject(new Error('detection blew up')));
await renderDialog(stub);
await waitForLocationHydrate();
+ await waitFor(() => {
+ expect(screen.getByTestId('create-editors-status').getAttribute('data-status')).toBe('none');
+ });
+ expect(screen.queryByTestId('create-editors-checkbox')).toBeNull();
+
await typeProjectName(PROJECT_NAME);
await waitForSubmitEnabled();
fireEvent.click(screen.getByTestId('create-submit'));
@@ -326,11 +503,9 @@ describe('CreateProjectDialog runtime wiring', () => {
expect(stub.createNewCalls[0]?.editors).toEqual([]);
});
- test('a detection probe that lands after the user ticks does not clobber the selection', async () => {
- // Seeding is an async round-trip the user can beat: Advanced is one click
- // away and the checkboxes are live well before `integrations.status()`
- // resolves. Whatever the checkboxes show after the user's click is what
- // gets submitted — a late seed neither replaces nor merges into it.
+ test('an in-flight probe shows a checking state, and a late result cannot flip the answer', async () => {
+ // The probe is a round-trip the user can beat. It fills the list of tools
+ // the label names; it never changes the answer the user already gave.
let releaseDetection = (): void => {};
const stub = makeBridge();
stub.setDetectedEditorsImpl(
@@ -342,26 +517,31 @@ describe('CreateProjectDialog runtime wiring', () => {
await renderDialog(stub);
await waitForLocationHydrate();
- fireEvent.click(screen.getByTestId('create-advanced-trigger'));
- fireEvent.click(screen.getByTestId(`create-editor-${UNDETECTED}`));
- expect(screen.getByTestId(`create-editor-${UNDETECTED}`).getAttribute('aria-checked')).toBe(
- 'true',
- );
+ // While probing: no checkbox to click yet, and the row says why. The status
+ // region is always mounted, so assert its state rather than its presence.
+ expect(screen.getByTestId('create-editors-status').getAttribute('data-status')).toBe('probing');
+ expect(screen.queryByTestId('create-editors-checkbox')).toBeNull();
releaseDetection();
+ await waitFor(() => {
+ expect(screen.queryByTestId('create-editors-checkbox')).not.toBeNull();
+ });
+ // Same node, new state — the region survived the transition rather than
+ // being torn down and remounted, which is what makes it announceable.
+ expect(screen.getByTestId('create-editors-status').getAttribute('data-status')).toBe('ready');
+ // Pre-checked on arrival; unchecking after the fact still wins.
+ fireEvent.click(screen.getByTestId('create-editors-checkbox'));
+ expect(screen.getByTestId('create-editors-checkbox').getAttribute('aria-checked')).toBe(
+ 'false',
+ );
await typeProjectName(PROJECT_NAME);
await waitForSubmitEnabled();
- expect(screen.getByTestId(`create-editor-${UNDETECTED}`).getAttribute('aria-checked')).toBe(
- 'true',
- );
- expect(screen.getByTestId('create-editor-claude').getAttribute('aria-checked')).toBe('false');
-
fireEvent.click(screen.getByTestId('create-submit'));
await waitFor(() => {
expect(stub.createNewCalls).toHaveLength(1);
});
- expect(stub.createNewCalls[0]?.editors).toEqual([UNDETECTED]);
+ expect(stub.createNewCalls[0]?.editors).toEqual([]);
});
test('reopening the dialog re-collapses Advanced so sharing is hidden again', async () => {
diff --git a/packages/app/src/components/CreateProjectDialog.tsx b/packages/app/src/components/CreateProjectDialog.tsx
index d1f17e267..635d92e75 100644
--- a/packages/app/src/components/CreateProjectDialog.tsx
+++ b/packages/app/src/components/CreateProjectDialog.tsx
@@ -8,11 +8,12 @@
* Layout: shadcn Dialog. The form leads with a Project name <Input> (focused
* on open) followed by a Location field (read-only display + Browse button
* that picks the PARENT directory). A live "Will be created at: …" caption
- * shows the resolved target before submit. The config-sharing posture
- * (side-by-side radio cards) and the editor-checkbox group (pre-checked from
- * `bridge.integrations.status().detectedEditorIds` on open) both collapse under
- * an "Advanced settings" section. Cancel +
- * Create footer. Create stays enabled with an empty name — a click then
+ * shows the resolved target before submit. Below that sits the AI-tool
+ * decision (`ProjectAiToolsField`) — one pre-checked checkbox naming the
+ * detected tools, always visible, because whether the project is reachable
+ * from the user's agents is not an advanced concern. Only the config-sharing
+ * posture (side-by-side radio cards) collapses under "Advanced settings".
+ * Cancel + Create footer. Create stays enabled with an empty name — a click then
* surfaces an "Enter a project name" toast (see onSubmit) rather than sitting
* disabled with no hint. The two fields (`location`, `name`) are the source of
* truth; the submit IPC takes `{ parent: location, name, ... }` with no
@@ -65,14 +66,17 @@
*/
import {
- ALL_EDITOR_IDS,
CREATE_NEW_PROJECT_FAILURE_REASONS,
type CreateNewBannerKind,
type CreateNewProjectFailureReason,
EDITOR_LABELS,
+ EDITOR_PROJECT_CONFIG_PATH,
+ EDITOR_PROJECT_SKILL_ROOT,
+ RESERVED_PROJECT_SKILL_NAME,
+ receivesProjectIntegrationWrite,
sanitizeFolderName,
} from '@inkeep/open-knowledge-core';
-import type { MessageDescriptor } from '@lingui/core';
+import { i18n, type MessageDescriptor } from '@lingui/core';
import { msg } from '@lingui/core/macro';
import { Trans, useLingui } from '@lingui/react/macro';
import { ArrowLeft, ChevronRight } from 'lucide-react';
@@ -108,6 +112,7 @@ import type {
} from '@/lib/desktop-bridge-types';
import { PACK_BLURBS } from '@/lib/pack-copy';
import { seedClient } from '@/lib/seed-client';
+import { formatToolList } from '@/lib/tool-list-format';
import { cn } from '@/lib/utils';
/**
@@ -357,15 +362,20 @@ export function CreateProjectDialog({
// Project name typed into the always-present <Input>. The creation target
// is `joinPathPreview(location, sanitizeFolderName(name))`.
const [name, setName] = useState('');
- // AI tools that get wired on submit. Seeded on each open from
- // `bridge.integrations.status().detectedEditorIds` — the tools whose host root
- // already exists on this machine — because the checkboxes live behind a collapsed
- // "Advanced settings" section: a user who types a name and clicks Create
- // never sees them, so an empty default would silently produce a project with
- // no MCP config and no project skill for tools they DO have. Seeding from
- // detection keeps the other half of that contract too: we never create a
- // host root for a tool the user doesn't have.
- const [editorIds, setEditorIds] = useState<ReadonlySet<OkMcpWiringEditorId>>(() => new Set());
+ // Tools detected on this machine that also have a project surface, probed on
+ // each open. `null` means the probe is still in flight — distinct from `[]`
+ // ("probed, found nothing"), because the row is always visible now and has to
+ // say which of the two it is rather than flashing an empty state.
+ const [detectedEditors, setDetectedEditors] = useState<readonly OkMcpWiringEditorId[] | null>(
+ null,
+ );
+ // Whether to wire those tools on submit. One decision, pre-checked: the write
+ // set is exactly the detected tools, so there is nothing to seed and no race
+ // with the probe — a late result changes the list the label names, never the
+ // answer the user gave.
+ const [connectEditors, setConnectEditors] = useState(true);
+ // Expands the exact list of files the connection writes.
+ const [showEditorDetails, setShowEditorDetails] = useState(false);
// OK config sharing mode. Defaults to `'shared'` (encourages team
// adoption). Rendered via SharingModeField inside "Advanced settings" — the
// greenfield dialog tucks the choice away (sensible default already set),
@@ -374,9 +384,8 @@ export function CreateProjectDialog({
// always runs `ensureProjectGit` (step 6 of runCreateNew), so the gitdir is
// guaranteed to exist by the time the sharing transition runs.
const [sharing, setSharing] = useState<'shared' | 'local-only'>('shared');
- // Editor controls and the sharing posture collapse under "Advanced settings"
- // so the dialog leads with just the name and location fields. Reset closed
- // on each open.
+ // The sharing posture collapses under "Advanced settings" so the dialog leads
+ // with just the name, location and AI-tool fields. Reset closed on each open.
const [advancedOpen, setAdvancedOpen] = useState(false);
// Starter-pack selection + where it scaffolds. `packId` starts at the
// launcher's pick and can be changed in-dialog (step 'pick'); the root
@@ -418,12 +427,6 @@ export function CreateProjectDialog({
// banner, so a result that arrives after the banner has moved on would
// otherwise be lost without UX feedback).
const removeGitCallIdRef = useRef(0);
- // True once the user has toggled an editor checkbox during the current open.
- // The on-open detection probe is async, so a user who expands Advanced and
- // ticks a box before it resolves would otherwise have their choice silently
- // overwritten by the seed landing late. What the checkboxes show right after
- // their click is the truth; a late probe never overrides it.
- const editorsTouchedRef = useRef(false);
// Whether the next pack-preview plan is the first of this open cycle — it
// fires immediately, later ones debounce. Reset on open and whenever the
// user picks a different pack.
@@ -445,8 +448,9 @@ export function CreateProjectDialog({
setProbeLifecycle('idle');
setBusy(false);
setName('');
- setEditorIds(new Set());
- editorsTouchedRef.current = false;
+ setDetectedEditors(null);
+ setConnectEditors(true);
+ setShowEditorDetails(false);
setSharing('shared');
setAdvancedOpen(false);
setRemoveGitState({ kind: 'idle' });
@@ -464,19 +468,42 @@ export function CreateProjectDialog({
let cancelled = false;
// Re-probe detection on every open (the user may have installed a tool
- // since last time). A selection the user has already edited wins over the
- // seed: this probe is a round-trip the user can beat.
+ // since last time). Filtered to the tools this create will actually write
+ // something for — `receivesProjectIntegrationWrite`, not mere surface
+ // membership. A user-global-only tool has nothing to write; Copilot has a
+ // project skill root but its skill is gated on Copilot's user-global
+ // OpenKnowledge entry, so before that exists the write lands as
+ // `skipped-prerequisite`. Naming either in the checkbox label would promise
+ // a file that never appears.
bridge.integrations
.status()
.then((status) => {
- if (!cancelled && !editorsTouchedRef.current) {
- setEditorIds(new Set(status.detectedEditorIds));
- }
+ if (cancelled) return;
+ // `installed` only — deliberately stricter than the write path's own
+ // check, which asks whether ANY entry sits under OpenKnowledge's server
+ // name and so also passes on `foreign` (an entry under that name that
+ // isn't ours). A foreign entry means OK's MCP is not actually
+ // registered, so the skill would tell the agent to call tools that
+ // aren't there. Erring strict costs a Copilot user with a foreign entry
+ // nothing they were promised; erring loose would name a tool whose
+ // setup does not work.
+ const userMcpInstalled = new Set(
+ status.editors.filter((e) => e.state === 'installed').map((e) => e.id),
+ );
+ setDetectedEditors(
+ status.detectedEditorIds.filter((id) =>
+ receivesProjectIntegrationWrite(id, {
+ userMcpEntryInstalled: userMcpInstalled.has(id),
+ }),
+ ),
+ );
})
.catch((err) => {
- // Best-effort: leave the selection empty so we never create a host
- // root for a tool we could not confirm. Advanced settings still works.
+ // Best-effort: settle on an empty list so we never create a host root
+ // for a tool we could not confirm, and the row says so rather than
+ // hanging on "Checking…".
console.warn('[CreateProjectDialog] editor-detection probe failed:', err);
+ if (!cancelled) setDetectedEditors([]);
});
// Reset Location before refetching — second-open after a first-open
@@ -551,9 +578,10 @@ export function CreateProjectDialog({
selectedPack !== undefined && rootChoice === 'subfolder' && trimmedSubfolder === '';
const packRootDir = rootChoice === 'project-root' ? undefined : trimmedSubfolder;
// The pack's skills install only into editors this project is set up for,
- // and `runCreateNew` writes those integrations before it seeds — so an empty
- // editor selection means no skill lands, and the preview must say so.
- const skillsInstallable = editorIds.size > 0;
+ // and `runCreateNew` writes those integrations before it seeds — so declining
+ // the AI-tool setup (or finding no tool to set up) means no skill lands, and
+ // the preview must say so.
+ const skillsInstallable = connectEditors && (detectedEditors?.length ?? 0) > 0;
// Live pack preview. Re-plans on every input that changes what would be
// written; nothing here touches disk (main plans against a throwaway dir).
@@ -837,17 +865,16 @@ export function CreateProjectDialog({
// ("Enter a project name", see onSubmit). Genuinely-blocked states
// (in-flight probe, blocking cascade, unusable name) stay disabled
// because they already render inline feedback that explains the block.
- const submitDisabled = busy || (rawName !== '' && !canSubmit);
-
- function toggleEditor(id: OkMcpWiringEditorId) {
- editorsTouchedRef.current = true;
- setEditorIds((prev) => {
- const next = new Set(prev);
- if (next.has(id)) next.delete(id);
- else next.add(id);
- return next;
- });
- }
+ //
+ // The detection probe is one of those blocks. It settles independently of the
+ // location/cascade probes, so Create can otherwise unlock while
+ // `detectedEditors` is still null — and submitting then sends `editors: []`,
+ // creating a project wired to nothing while the row still says "Checking
+ // which AI tools you have". Only gate it while the user actually intends to
+ // connect: with the box unchecked the list is never read, so there is nothing
+ // to wait for. The row's own status text explains the wait.
+ const detectionPending = connectEditors && detectedEditors === null;
+ const submitDisabled = busy || detectionPending || (rawName !== '' && !canSubmit);
async function onBrowse() {
try {
@@ -900,7 +927,7 @@ export function CreateProjectDialog({
await bridge.project.createNew({
parent: location,
name: sanitized,
- editors: Array.from(editorIds),
+ editors: connectEditors ? [...(detectedEditors ?? [])] : [],
sharing,
// Seed the chosen starter pack (packs-forward first-run). Undefined on
// the blank-create path — main opens an empty project as before.
@@ -1207,6 +1234,20 @@ export function CreateProjectDialog({
</div>
) : null}
+ {/* AI-tool setup, always visible: it decides whether the project is
+ usable from the user's agents at all, which is not an advanced
+ concern. Mirrors the first-launch consent dialog — one pre-checked
+ checkbox whose label names the write set, plus an expander with the
+ exact files. Per-tool control lives in Settings > This project. */}
+ <ProjectAiToolsField
+ detectedEditors={detectedEditors}
+ checked={connectEditors}
+ onCheckedChange={setConnectEditors}
+ showDetails={showEditorDetails}
+ onShowDetailsChange={setShowEditorDetails}
+ disabled={busy}
+ />
+
<Collapsible
open={advancedOpen}
onOpenChange={setAdvancedOpen}
@@ -1224,30 +1265,6 @@ export function CreateProjectDialog({
/>
</CollapsibleTrigger>
<CollapsibleContent className="space-y-6 border-t border-border px-3 py-4">
- <fieldset className="flex flex-col space-y-2 pb-2">
- <legend className="text-sm font-medium">
- <Trans>Connect to AI tools</Trans>
- </legend>
- <p className="text-1sm text-muted-foreground">
- <Trans>Each selected tool gets an OpenKnowledge MCP entry.</Trans>
- </p>
- {ALL_EDITOR_IDS.map((id) => {
- const inputId = `create-editor-${id}`;
- return (
- <Label key={id} htmlFor={inputId} className="text-sm font-normal">
- <Checkbox
- id={inputId}
- checked={editorIds.has(id)}
- onCheckedChange={() => toggleEditor(id)}
- disabled={busy}
- data-testid={`create-editor-${id}`}
- />
- <span>{EDITOR_LABELS[id]}</span>
- </Label>
- );
- })}
- </fieldset>
-
<SharingModeField
idPrefix="create"
testIdPrefix="create-sharing"
@@ -1311,6 +1328,177 @@ export function CreateProjectDialog({
);
}
+interface ProjectAiToolsFieldProps {
+ /** `null` while the detection probe is in flight; `[]` once it settled empty. */
+ detectedEditors: readonly OkMcpWiringEditorId[] | null;
+ checked: boolean;
+ onCheckedChange: (next: boolean) => void;
+ showDetails: boolean;
+ onShowDetailsChange: (next: boolean) => void;
+ disabled: boolean;
+}
+
+/**
+ * The project's AI-tool decision: one checkbox covering every detected tool,
+ * with a read-only expander naming the exact project-relative files it writes.
+ *
+ * The paths come from the same two core maps the project writer resolves its
+ * targets from (`EDITOR_PROJECT_CONFIG_PATH`, `EDITOR_PROJECT_SKILL_ROOT`), so
+ * the disclosure cannot advertise a file that never gets written. Tools with a
+ * null entry in both are filtered out upstream and never reach this component.
+ */
+function ProjectAiToolsField({
+ detectedEditors,
+ checked,
+ onCheckedChange,
+ showDetails,
+ onShowDetailsChange,
+ disabled,
+}: ProjectAiToolsFieldProps) {
+ const { t } = useLingui();
+ const detailsId = useId();
+ const checkboxId = useId();
+
+ // The probe's two non-interactive outcomes share ONE live region that is
+ // always mounted. A region that appears at the same moment as its text is not
+ // a change to announce — assistive tech has to be observing the node before
+ // the content lands — so swapping the message inside a persistent node is what
+ // makes "checking…" → "none found" actually reach a screen reader.
+ const status =
+ detectedEditors === null
+ ? { kind: 'probing' as const, message: t`Checking which AI tools you have` }
+ : detectedEditors.length === 0
+ ? {
+ kind: 'none' as const,
+ message: t`No AI tools detected yet. Once you install one, connect it from Settings > This project.`,
+ }
+ : // Carrying the narrowed list on the ready arm is what lets the early
+ // return below discriminate it — the guard alone doesn't re-narrow
+ // `detectedEditors` for the JSX that follows.
+ { kind: 'ready' as const, message: '', editors: detectedEditors };
+
+ const statusRegion = (
+ <p
+ aria-live="polite"
+ className={cn(
+ status.kind === 'ready'
+ ? 'sr-only'
+ : 'rounded-md border border-border px-3 py-2.5 text-1sm text-muted-foreground',
+ )}
+ data-status={status.kind}
+ data-testid="create-editors-status"
+ >
+ {status.message}
+ </p>
+ );
+
+ if (status.kind !== 'ready') return statusRegion;
+
+ return (
+ <>
+ {/* Stays mounted (visually hidden, empty) so the region survives every
+ transition rather than being torn down when the checkbox appears. */}
+ {statusRegion}
+ <div className="overflow-hidden rounded-md border border-border">
+ <Label
+ htmlFor={checkboxId}
+ className="flex cursor-pointer items-start gap-2.5 px-3 py-2.5 font-normal hover:bg-muted/50"
+ >
+ <Checkbox
+ id={checkboxId}
+ checked={checked}
+ onCheckedChange={() => onCheckedChange(!checked)}
+ disabled={disabled}
+ className="mt-0.5"
+ data-testid="create-editors-checkbox"
+ />
+ <span className="flex min-w-0 flex-1 flex-col gap-0.5">
+ <span
+ className="text-sm font-medium text-foreground"
+ data-testid="create-editors-summary"
+ >
+ {t`Connect ${formatToolList(
+ status.editors.map((id) => EDITOR_LABELS[id]),
+ i18n.locale,
+ )} to this project`}
+ </span>
+ <span className="text-1sm text-muted-foreground">
+ <Trans comment="Subtext under the create-project AI-tools checkbox">
+ Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and
+ update these notes.
+ </Trans>
+ </span>
+ </span>
+ </Label>
+ <div className="border-t border-border">
+ <Button
+ type="button"
+ variant="ghost"
+ size="sm"
+ className="h-auto w-full justify-start gap-1 rounded-none px-3 py-2 text-xs font-normal text-muted-foreground hover:text-foreground"
+ onClick={() => onShowDetailsChange(!showDetails)}
+ aria-expanded={showDetails}
+ aria-controls={detailsId}
+ data-testid="create-editors-details-toggle"
+ >
+ <ChevronRight
+ className={cn(
+ 'size-3.5 transition-transform motion-reduce:transition-none',
+ showDetails && 'rotate-90',
+ )}
+ aria-hidden
+ />
+ {showDetails ? (
+ <Trans comment="Collapses the list of project files the AI-tool setup writes">
+ Hide details
+ </Trans>
+ ) : (
+ <Trans comment="Expands the list of project files the AI-tool setup writes">
+ What this changes
+ </Trans>
+ )}
+ </Button>
+ {/* The region element is always present so the toggle's `aria-controls`
+ never dangles — axe flags a reference to an id that isn't in the DOM,
+ which is what a conditionally-mounted target produces. */}
+ <div id={detailsId}>
+ {showDetails && (
+ <ul
+ className="flex flex-col gap-1.5 border-t border-border px-3 py-2.5"
+ data-testid="create-editors-details"
+ >
+ {status.editors.map((id) => {
+ const configPath = EDITOR_PROJECT_CONFIG_PATH[id];
+ const skillRoot = EDITOR_PROJECT_SKILL_ROOT[id];
+ return (
+ <li
+ key={id}
+ className="flex min-w-0 flex-col"
+ data-testid={`create-editor-${id}`}
+ >
+ <span className="text-1sm text-foreground">{EDITOR_LABELS[id]}</span>
+ {configPath !== null && (
+ <code className="text-xs text-muted-foreground break-all">
+ {configPath}
+ </code>
+ )}
+ {skillRoot !== null && (
+ <code className="text-xs text-muted-foreground break-all">
+ {`${skillRoot}/${RESERVED_PROJECT_SKILL_NAME}/`}
+ </code>
+ )}
+ </li>
+ );
+ })}
+ </ul>
+ )}
+ </div>
+ </div>
+ </div>
+ </>
+ );
+}
+
interface CascadeBannerProps {
cascade: SettledCascade;
onOpenNested: (rootPath: string) => void;
diff --git a/packages/app/src/components/McpConsentDialog.dom.test.tsx b/packages/app/src/components/McpConsentDialog.dom.test.tsx
index e22ba4574..3362e2888 100644
--- a/packages/app/src/components/McpConsentDialog.dom.test.tsx
+++ b/packages/app/src/components/McpConsentDialog.dom.test.tsx
@@ -1,5 +1,5 @@
import * as actualLinguiMacro from '@lingui/react/macro';
-import { cleanup, render, screen, waitFor, within } from '@testing-library/react';
+import { cleanup, render, screen, waitFor } from '@testing-library/react';
import userEvent from '@testing-library/user-event';
import type { ReactNode } from 'react';
import { afterEach, describe, expect, test, vi } from 'vitest';
@@ -14,6 +14,12 @@ vi.doMock('@lingui/react/macro', () => ({
useLingui: () => ({ t: renderLinguiTemplate }),
}));
+const DISCOVERY_SKILL = {
+ id: 'discovery',
+ name: 'open-knowledge-discovery',
+ paths: ['~/.agents/skills/open-knowledge-discovery', '~/.claude/skills/open-knowledge-discovery'],
+};
+
const payload: OkMcpWiringShowPayload = {
detectedEditors: [
{
@@ -46,10 +52,10 @@ const payload: OkMcpWiringShowPayload = {
rcFilesToTouch: ['~/.zshrc', '~/.config/fish/conf.d/open-knowledge.fish'],
alreadyInstalled: false,
},
- globalSkills: [],
+ globalSkills: [DISCOVERY_SKILL],
};
-/** Same editors, zero detected — exercises the Add-enable matrix. */
+/** Same shell state, zero detected tools — exercises the empty state. */
const noneDetectedPayload: OkMcpWiringShowPayload = {
detectedEditors: [
{
@@ -62,7 +68,7 @@ const noneDetectedPayload: OkMcpWiringShowPayload = {
},
],
pathInstall: payload.pathInstall,
- globalSkills: [],
+ globalSkills: [DISCOVERY_SKILL],
};
function deferredResult() {
@@ -76,57 +82,9 @@ function deferredResult() {
interface RecordedConfirm {
editorIds: readonly string[];
pathInstall: boolean | undefined;
- skills?: readonly string[];
+ skills: readonly string[] | undefined;
}
-/** Payload variant that offers both skill rows — exercises the skills section. */
-const skillsPayload: OkMcpWiringShowPayload = {
- detectedEditors: [
- {
- id: 'claude',
- label: 'Claude',
- detected: true,
- willReplace: false,
- configPath: '~/.claude.json',
- entryLocator: 'mcpServers.open-knowledge',
- },
- ],
- pathInstall: { shellDetected: false, rcFilesToTouch: [], alreadyInstalled: false },
- globalSkills: [
- {
- id: 'discovery',
- name: 'open-knowledge-discovery',
- alreadyInstalled: false,
- description: 'Recognize OpenKnowledge projects and route reads and writes through it.',
- size: { alwaysOn: 140, onTrigger: 1495, onDemand: 0 },
- hosts: ['claude', 'cursor'],
- paths: [
- '~/.agents/skills/open-knowledge-discovery',
- '~/.claude/skills/open-knowledge-discovery',
- ],
- },
- {
- id: 'write-skill',
- name: 'open-knowledge-write-skill',
- alreadyInstalled: true,
- description: 'A guided workflow for authoring new Agent Skills.',
- size: { alwaysOn: 156, onTrigger: 3218, onDemand: 916 },
- hosts: ['claude', 'cursor'],
- paths: [
- '~/.agents/skills/open-knowledge-write-skill',
- '~/.claude/skills/open-knowledge-write-skill',
- ],
- },
- ],
-};
-
-/** Same two bundles, zero resolved hosts — exercises the disabled/unchecked
- * state and the row's no-hosts copy (nothing can land). */
-const skillsNoHostsPayload: OkMcpWiringShowPayload = {
- ...skillsPayload,
- globalSkills: skillsPayload.globalSkills.map((s) => ({ ...s, hosts: [], paths: [] })),
-};
-
function makeHarness({
confirmResult = async () => ({ ok: true as const }),
skipResult = async () => ({ ok: true as const }),
@@ -145,7 +103,7 @@ function makeHarness({
confirmCalls.push({
editorIds: [...request.editorIds],
pathInstall: request.pathInstall,
- skills: request.skills ? [...request.skills] : undefined,
+ skills: request.skills ? [...request.skills] : request.skills,
});
return confirmResult(request.editorIds);
},
@@ -169,8 +127,8 @@ async function renderDialog(harness = makeHarness()) {
const { McpConsentDialogBody } = await import('./McpConsentDialogBody');
const { TooltipProvider } = await import('@/components/ui/tooltip');
// Mirror production: the app mounts a single root TooltipProvider (main.tsx),
- // and the dialog's per-row info tooltips rely on it rather than each wrapping
- // their own — so the isolated render must supply it too.
+ // and the PATH row's info tooltip relies on it rather than wrapping its own —
+ // so the isolated render must supply it too.
render(
<TooltipProvider>
<McpConsentDialogBody
@@ -183,108 +141,72 @@ async function renderDialog(harness = makeHarness()) {
return harness;
}
-describe('McpConsentDialog runtime behavior', () => {
+describe('McpConsentDialog AI-tools decision', () => {
afterEach(() => cleanup());
- test('renders willReplace disclosure and preselects detected editors only', async () => {
+ test('one pre-checked box whose label names every tool in the write set', async () => {
await renderDialog();
expect(
screen.getByRole('dialog', { name: 'Connect your AI tools to OpenKnowledge' }),
).toBeTruthy();
- expect(screen.getByTestId('mcp-consent-status-claude').textContent).toBe(
- 'Will replace existing OpenKnowledge entry',
- );
- // Detected tools carry no status line — the checked box conveys it.
- expect(screen.queryByTestId('mcp-consent-status-cursor')).toBeNull();
- // Progressive disclosure: undetected codex starts hidden behind the toggle.
- expect(screen.queryByTestId('mcp-consent-checkbox-codex')).toBeNull();
- expect(screen.getByTestId('mcp-consent-checkbox-claude').getAttribute('aria-checked')).toBe(
- 'true',
- );
- expect(screen.getByTestId('mcp-consent-checkbox-cursor').getAttribute('aria-checked')).toBe(
+ expect(screen.getByTestId('mcp-consent-connect-checkbox').getAttribute('aria-checked')).toBe(
'true',
);
+ // Consent integrity: collapsed, the summary still discloses the write set.
+ const summary = screen.getByTestId('mcp-consent-connect-summary').textContent ?? '';
+ expect(summary).toContain('Claude');
+ expect(summary).toContain('Cursor');
+ // Undetected tools are not in the write set and are not named.
+ expect(summary).not.toContain('Codex');
+ });
- // Reveal the undetected tools — codex links to its setup guide, unchecked.
- await userEvent.click(screen.getByTestId('mcp-consent-editors-toggle'));
- const codexStatus = screen.getByTestId('mcp-consent-status-codex');
- expect(codexStatus.tagName).toBe('A');
- expect(codexStatus.textContent).toContain('How to set up');
- expect(codexStatus.getAttribute('href')).toBe(
- 'https://openknowledge.ai/docs/integrations/codex',
- );
- expect(screen.getByTestId('mcp-consent-checkbox-codex').getAttribute('aria-checked')).toBe(
- 'false',
- );
+ test('consent integrity: the overwrite warning shows without expanding anything', async () => {
+ await renderDialog();
+
+ // Claude carries willReplace; the warning must be visible while the
+ // disclosure is still collapsed, naming the tool whose entry is replaced.
+ expect(screen.queryByTestId('mcp-consent-details')).toBeNull();
+ const warning = screen.getByTestId('mcp-consent-connect-replace-warning').textContent ?? '';
+ expect(warning).toContain('Claude');
+ expect(warning).not.toContain('Cursor');
});
- test('consent integrity: a checked undetected tool stays visible after collapse', async () => {
- // Two undetected tools so a toggle still exists after one is checked.
+ test('no overwrite warning when nothing will be replaced', async () => {
await renderDialog(
makeHarness({
snapshot: {
- detectedEditors: [
- {
- id: 'claude',
- label: 'Claude',
- detected: true,
- willReplace: false,
- configPath: '~/.claude.json',
- entryLocator: 'mcpServers.open-knowledge',
- },
- {
- id: 'codex',
- label: 'Codex',
- detected: false,
- willReplace: false,
- configPath: '~/.codex/config.toml',
- entryLocator: '[mcp_servers.open-knowledge]',
- },
- {
- id: 'opencode',
- label: 'OpenCode',
- detected: false,
- willReplace: false,
- configPath: '~/.config/opencode/opencode.json',
- entryLocator: 'mcp.open-knowledge',
- },
- ],
- pathInstall: { shellDetected: false, rcFilesToTouch: [], alreadyInstalled: false },
- globalSkills: [],
+ ...payload,
+ detectedEditors: payload.detectedEditors.map((e) => ({ ...e, willReplace: false })),
},
}),
);
- // Both undetected tools hidden until expanded.
- expect(screen.queryByTestId('mcp-consent-checkbox-codex')).toBeNull();
-
- await userEvent.click(screen.getByTestId('mcp-consent-editors-toggle'));
- await userEvent.click(screen.getByTestId('mcp-consent-checkbox-codex'));
- expect(screen.getByTestId('mcp-consent-checkbox-codex').getAttribute('aria-checked')).toBe(
- 'true',
- );
-
- // Collapse: checked codex stays visible (still in the write set); unchecked
- // opencode hides again, so the toggle persists.
- await userEvent.click(screen.getByTestId('mcp-consent-editors-toggle'));
- expect(screen.getByTestId('mcp-consent-checkbox-codex').getAttribute('aria-checked')).toBe(
- 'true',
- );
- expect(screen.queryByTestId('mcp-consent-checkbox-opencode')).toBeNull();
+ expect(screen.queryByTestId('mcp-consent-connect-replace-warning')).toBeNull();
});
- test('location tooltip discloses the config file and entry locator on focus', async () => {
+ test('the disclosure lists every config file, entry and skill destination', async () => {
await renderDialog();
- // Content is portaled — mounts only once the info trigger is focused. Radix
- // renders it twice when open (visible + a11y mirror), so assert getAllByText.
- screen.getByTestId('mcp-consent-editor-info-claude').focus();
- await waitFor(() => {
- expect(screen.getAllByText('~/.claude.json').length).toBeGreaterThan(0);
- expect(screen.getAllByText('mcpServers.open-knowledge').length).toBeGreaterThan(0);
- });
+
+ await userEvent.click(screen.getByTestId('mcp-consent-details-toggle'));
+ const details = screen.getByTestId('mcp-consent-details');
+
+ expect(details.textContent).toContain('~/.claude.json');
+ expect(details.textContent).toContain('mcpServers.open-knowledge');
+ expect(details.textContent).toContain('~/.cursor/mcp.json');
+ // Undetected tools get no row — nothing is written for them.
+ expect(screen.queryByTestId('mcp-consent-detail-codex')).toBeNull();
+
+ // Skill destinations come from the payload (main computes them from the
+ // installer's own gates), never re-derived in the renderer.
+ const skillDetail = screen.getByTestId('mcp-consent-detail-skill-discovery');
+ expect(skillDetail.textContent).toContain('~/.agents/skills/open-knowledge-discovery');
+ expect(skillDetail.textContent).toContain('~/.claude/skills/open-knowledge-discovery');
+
+ await userEvent.click(screen.getByTestId('mcp-consent-details-toggle'));
+ expect(screen.queryByTestId('mcp-consent-details')).toBeNull();
});
- test('location tooltip shows the null-configPath fallback', async () => {
+ test('the null-configPath fallback renders in the disclosure', async () => {
// claude-desktop has no user-global config on this platform (configPath null).
await renderDialog(
makeHarness({
@@ -293,7 +215,7 @@ describe('McpConsentDialog runtime behavior', () => {
{
id: 'claude-desktop',
label: 'Claude Desktop',
- detected: false,
+ detected: true,
willReplace: false,
configPath: null,
entryLocator: 'mcpServers.open-knowledge',
@@ -304,122 +226,131 @@ describe('McpConsentDialog runtime behavior', () => {
},
}),
);
- screen.getByTestId('mcp-consent-editor-info-claude-desktop').focus();
+
+ await userEvent.click(screen.getByTestId('mcp-consent-details-toggle'));
+ expect(screen.getByTestId('mcp-consent-detail-claude-desktop').textContent).toContain(
+ 'unavailable on this platform',
+ );
+ });
+
+ test('Continue sends every detected tool plus the offered skill bundles', async () => {
+ const harness = await renderDialog();
+
+ await userEvent.click(screen.getByTestId('mcp-consent-add'));
await waitFor(() => {
- expect(screen.getAllByText('unavailable on this platform').length).toBeGreaterThan(0);
+ expect(harness.confirmCalls).toEqual([
+ { editorIds: ['claude', 'cursor'], pathInstall: true, skills: ['discovery'] },
+ ]);
});
});
- test('undetected claude-desktop links to the shared claude-code guide (aliased slug)', async () => {
- // claude-desktop → claude-code is the only non-1:1 entry in
- // EDITOR_SETUP_DOC_SLUG; a regression to `editor.id` in the URL would 404.
- await renderDialog(
- makeHarness({
- snapshot: {
- detectedEditors: [
- {
- id: 'claude-desktop',
- label: 'Claude Desktop',
- detected: false,
- willReplace: false,
- configPath: null,
- entryLocator: 'mcpServers.open-knowledge',
- },
- ],
- pathInstall: { shellDetected: false, rcFilesToTouch: [], alreadyInstalled: false },
- globalSkills: [],
- },
- }),
- );
+ test('unchecking sends no editors AND no skill decision — declining never removes', async () => {
+ const harness = await renderDialog();
- const status = screen.getByTestId('mcp-consent-status-claude-desktop');
- expect(status.tagName).toBe('A');
- expect(status.getAttribute('href')).toBe(
- 'https://openknowledge.ai/docs/integrations/claude-code',
- );
+ await userEvent.click(screen.getByTestId('mcp-consent-connect-checkbox'));
+ await userEvent.click(screen.getByTestId('mcp-consent-add'));
+
+ await waitFor(() => {
+ // `skills: undefined`, not `[]`: main reads an array as "decline every
+ // offered bundle and tear down any that is installed", which this screen
+ // must never do.
+ expect(harness.confirmCalls).toEqual([
+ { editorIds: [], pathInstall: true, skills: undefined },
+ ]);
+ });
+ expect(harness.toastMessages).toEqual(['This can be configured in Settings > AI tools & CLI']);
});
- test('failed Add resets busy state, reports the error, and allows retry', async () => {
- const first = deferredResult();
- const second = deferredResult();
- const outcomes = [first, second];
- const harness = makeHarness({
- confirmResult: async () => outcomes.shift()?.promise ?? { ok: true },
+ test('connecting does not fire the Settings pointer toast', async () => {
+ const harness = await renderDialog();
+ await userEvent.click(screen.getByTestId('mcp-consent-add'));
+ await waitFor(() => {
+ expect(harness.confirmCalls.length).toBe(1);
});
- await renderDialog(harness);
+ expect(harness.toastMessages).toEqual([]);
+ });
- const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
- const skip = screen.getByTestId('mcp-consent-skip') as HTMLButtonElement;
+ test('no skills offered: subtext claims only MCP, and confirm sends no skill decision', async () => {
+ // `skillsOffered = false` drives two things a consent screen must not get
+ // wrong: the subtext must stop promising the discovery skill, and the
+ // confirm must send `skills: undefined` rather than an array — an array
+ // would record a decline for bundles that were never offered.
+ const harness = await renderDialog(makeHarness({ snapshot: { ...payload, globalSkills: [] } }));
- await userEvent.click(add);
- expect(add.disabled).toBe(true);
- expect(skip.disabled).toBe(true);
- expect(add.textContent).toBe('Working');
+ const row = screen.getByTestId('mcp-consent-connect-checkbox').closest('label');
+ expect(row?.textContent ?? '').not.toContain('discovery');
+ expect(row?.textContent ?? '').not.toContain('skill');
- first.resolve({ ok: false, error: 'Could not write Claude config' });
+ await userEvent.click(screen.getByTestId('mcp-consent-add'));
await waitFor(() => {
- expect(add.disabled).toBe(false);
+ expect(harness.confirmCalls).toEqual([
+ { editorIds: ['claude', 'cursor'], pathInstall: true, skills: undefined },
+ ]);
});
+ });
- expect(skip.disabled).toBe(false);
- expect(add.textContent).toBe('Connect');
- expect(harness.confirmCalls).toEqual([{ editorIds: ['claude', 'cursor'], pathInstall: true }]);
- expect(harness.toastErrors).toEqual(['Could not write Claude config']);
+ test('the overwrite warning disappears when the box is unchecked', async () => {
+ // "Replaces …" is present tense. Leaving it up after an uncheck states a
+ // consequence that will not happen, and reads as the uncheck not taking.
+ await renderDialog();
+ expect(screen.getByTestId('mcp-consent-connect-replace-warning')).toBeTruthy();
+ await userEvent.click(screen.getByTestId('mcp-consent-connect-checkbox'));
+ expect(screen.queryByTestId('mcp-consent-connect-replace-warning')).toBeNull();
+ });
+
+ test('no detected tools: no checkbox, an explanatory line, and a PATH-only confirm', async () => {
+ const harness = await renderDialog(makeHarness({ snapshot: noneDetectedPayload }));
+ expect(screen.queryByTestId('mcp-consent-connect-checkbox')).toBeNull();
+ expect(screen.getByTestId('mcp-consent-no-tools').textContent).toContain(
+ 'No AI tools detected',
+ );
+
+ const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
+ expect(add.disabled).toBe(false);
await userEvent.click(add);
- second.resolve({ ok: false, error: 'Still unwritable' });
await waitFor(() => {
expect(harness.confirmCalls).toEqual([
- { editorIds: ['claude', 'cursor'], pathInstall: true },
- { editorIds: ['claude', 'cursor'], pathInstall: true },
+ { editorIds: [], pathInstall: true, skills: undefined },
]);
});
});
- test('failed Skip resets busy state, reports the error, and allows retry', async () => {
+ test('Continue stays enabled with nothing selected — it always records a decision', async () => {
+ await renderDialog();
+ await userEvent.click(screen.getByTestId('mcp-consent-connect-checkbox'));
+ await userEvent.click(screen.getByTestId('mcp-consent-path-checkbox'));
+ expect((screen.getByTestId('mcp-consent-add') as HTMLButtonElement).disabled).toBe(false);
+ });
+
+ test('failed Continue resets busy state, reports the error, and allows retry', async () => {
const first = deferredResult();
const second = deferredResult();
const outcomes = [first, second];
const harness = makeHarness({
- skipResult: async () => outcomes.shift()?.promise ?? { ok: true },
+ confirmResult: async () => outcomes.shift()?.promise ?? { ok: true },
});
await renderDialog(harness);
const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
- const skip = screen.getByTestId('mcp-consent-skip') as HTMLButtonElement;
- await userEvent.click(skip);
+ await userEvent.click(add);
expect(add.disabled).toBe(true);
- expect(skip.disabled).toBe(true);
-
- first.resolve({ ok: false, error: 'Could not write marker' });
- await waitFor(() => {
- expect(skip.disabled).toBe(false);
- });
-
- expect(add.disabled).toBe(false);
- expect(harness.skipCalls).toEqual(['skip']);
- expect(harness.toastErrors).toEqual(['Could not write marker']);
+ expect(add.textContent).toBe('Working');
- await userEvent.click(skip);
- second.resolve({ ok: false, error: 'Still cannot write marker' });
+ first.resolve({ ok: false, error: 'Could not write Claude config' });
await waitFor(() => {
- expect(harness.skipCalls).toEqual(['skip', 'skip']);
+ expect(add.disabled).toBe(false);
});
- });
-
- test('successful Skip points the user at the Settings surface via a toast', async () => {
- const harness = makeHarness();
- await renderDialog(harness);
- await userEvent.click(screen.getByTestId('mcp-consent-skip'));
+ expect(add.textContent).toBe('Continue');
+ expect(harness.toastErrors).toEqual(['Could not write Claude config']);
+ await userEvent.click(add);
+ second.resolve({ ok: false, error: 'Still unwritable' });
await waitFor(() => {
- expect(harness.toastMessages).toEqual([
- 'This can be configured in Settings > AI tools & CLI',
- ]);
+ expect(harness.confirmCalls.length).toBe(2);
});
- expect(harness.toastErrors).toEqual([]);
});
});
@@ -456,7 +387,7 @@ describe('McpConsentDialog PATH consent row', () => {
);
});
- test('unchecking the toggle sends pathInstall:false on Add', async () => {
+ test('unchecking the toggle sends pathInstall:false on Continue', async () => {
const harness = await renderDialog();
await userEvent.click(screen.getByTestId('mcp-consent-path-checkbox'));
@@ -464,31 +395,11 @@ describe('McpConsentDialog PATH consent row', () => {
await waitFor(() => {
expect(harness.confirmCalls).toEqual([
- { editorIds: ['claude', 'cursor'], pathInstall: false },
+ { editorIds: ['claude', 'cursor'], pathInstall: false, skills: ['discovery'] },
]);
});
});
- test('FR8: zero editors selected + PATH checked keeps Add enabled and confirms PATH-only', async () => {
- const harness = await renderDialog(makeHarness({ snapshot: noneDetectedPayload }));
-
- const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
- expect(add.disabled).toBe(false);
-
- await userEvent.click(add);
- await waitFor(() => {
- expect(harness.confirmCalls).toEqual([{ editorIds: [], pathInstall: true }]);
- });
- });
-
- test('FR8: zero editors + PATH unchecked disables Add', async () => {
- await renderDialog(makeHarness({ snapshot: noneDetectedPayload }));
-
- await userEvent.click(screen.getByTestId('mcp-consent-path-checkbox'));
- const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
- expect(add.disabled).toBe(true);
- });
-
test('alreadyInstalled renders an informational row and solicits no decision', async () => {
const harness = await renderDialog(
makeHarness({
@@ -509,7 +420,7 @@ describe('McpConsentDialog PATH consent row', () => {
await userEvent.click(screen.getByTestId('mcp-consent-add'));
await waitFor(() => {
expect(harness.confirmCalls).toEqual([
- { editorIds: ['claude', 'cursor'], pathInstall: undefined },
+ { editorIds: ['claude', 'cursor'], pathInstall: undefined, skills: ['discovery'] },
]);
});
});
@@ -529,121 +440,42 @@ describe('McpConsentDialog PATH consent row', () => {
await userEvent.click(screen.getByTestId('mcp-consent-add'));
await waitFor(() => {
expect(harness.confirmCalls).toEqual([
- { editorIds: ['claude', 'cursor'], pathInstall: undefined },
+ { editorIds: ['claude', 'cursor'], pathInstall: undefined, skills: ['discovery'] },
]);
});
});
});
-describe('McpConsentDialog skills section', () => {
+describe('McpConsentDialog dismissal', () => {
afterEach(() => cleanup());
- test('FR2: renders one pre-checked row per bundle', async () => {
- await renderDialog(makeHarness({ snapshot: skillsPayload }));
- for (const id of ['discovery', 'write-skill']) {
- expect(
- screen.getByTestId(`mcp-consent-skill-checkbox-${id}`).getAttribute('aria-checked'),
- ).toBe('true');
- }
- });
+ test('Escape skips without recording any decision and points at Settings', async () => {
+ const harness = await renderDialog();
- test('FR9: unchecking write-skill sends only the checked bundle on Add', async () => {
- const harness = await renderDialog(makeHarness({ snapshot: skillsPayload }));
- await userEvent.click(screen.getByTestId('mcp-consent-skill-checkbox-write-skill'));
- // Unchecking an already-installed bundle surfaces the removal warning.
- expect(screen.getByTestId('mcp-consent-skill-warning-write-skill')).toBeTruthy();
- await userEvent.click(screen.getByTestId('mcp-consent-add'));
- await waitFor(() => {
- expect(harness.confirmCalls).toEqual([
- { editorIds: ['claude'], pathInstall: undefined, skills: ['discovery'] },
- ]);
- });
- });
+ await userEvent.keyboard('{Escape}');
- test('declining every skill keeps Add enabled and confirms an empty skill set', async () => {
- const harness = await renderDialog(makeHarness({ snapshot: skillsPayload }));
- await userEvent.click(screen.getByTestId('mcp-consent-skill-checkbox-discovery'));
- await userEvent.click(screen.getByTestId('mcp-consent-skill-checkbox-write-skill'));
- // Uncheck the only detected editor too — Add stays enabled because skills
- // were offered (declining is itself an action).
- await userEvent.click(screen.getByTestId('mcp-consent-checkbox-claude'));
- const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
- expect(add.disabled).toBe(false);
- await userEvent.click(add);
await waitFor(() => {
- expect(harness.confirmCalls).toEqual([{ editorIds: [], pathInstall: undefined, skills: [] }]);
+ expect(harness.skipCalls).toEqual(['skip']);
});
+ expect(harness.confirmCalls).toEqual([]);
+ expect(harness.toastMessages).toEqual(['This can be configured in Settings > AI tools & CLI']);
});
- test("renders each skill's own frontmatter description, not hand-written subtext", async () => {
- await renderDialog(makeHarness({ snapshot: skillsPayload }));
- expect(screen.getByText('A guided workflow for authoring new Agent Skills.')).toBeTruthy();
- });
-
- test('states skill reach is independent of the MCP editor selection', async () => {
- await renderDialog(makeHarness({ snapshot: skillsPayload }));
- expect(screen.getByTestId('mcp-consent-skill-fanout-note').textContent).toContain(
- 'independent of the MCP',
- );
- });
-
- test('clicking a row expands it in place with the full cost and destination list', async () => {
- await renderDialog(makeHarness({ snapshot: skillsPayload }));
- // Collapsed: no expansion mounted.
- expect(screen.queryByTestId('mcp-consent-skill-expansion-write-skill')).toBeNull();
-
- // The row body is the expand affordance (first launch expands; it never
- // navigates away). Clicking the skill name bubbles to its row button.
- await userEvent.click(screen.getByText('open-knowledge-write-skill'));
-
- const expansion = screen.getByTestId('mcp-consent-skill-expansion-write-skill');
- // Full three-tier cost — the on-demand tier appears only here, never on the
- // compact row above it.
- expect(expansion.textContent).toContain('on demand');
- // Every destination path is disclosed, including the ~/.agents hub — the
- // same list the Settings confirm modal carries.
- const paths = within(expansion).getByTestId('skill-destination-list');
- expect(paths.textContent).toContain('~/.agents/skills/open-knowledge-write-skill');
- expect(paths.textContent).toContain('~/.claude/skills/open-knowledge-write-skill');
- // The dialog is never left.
- expect(
- screen.getByRole('dialog', { name: 'Connect your AI tools to OpenKnowledge' }),
- ).toBeTruthy();
- });
+ test('failed skip resets busy state and reports the error', async () => {
+ const first = deferredResult();
+ const harness = makeHarness({ skipResult: async () => first.promise });
+ await renderDialog(harness);
- test('checking a skill stages it — nothing is written until Connect', async () => {
- const harness = await renderDialog(makeHarness({ snapshot: skillsPayload }));
- // Toggle discovery off then on — the checkbox itself never confirms.
- await userEvent.click(screen.getByTestId('mcp-consent-skill-checkbox-discovery'));
- await userEvent.click(screen.getByTestId('mcp-consent-skill-checkbox-discovery'));
- expect(harness.confirmCalls).toEqual([]);
- // Only Connect commits the staged set.
- await userEvent.click(screen.getByTestId('mcp-consent-add'));
+ await userEvent.keyboard('{Escape}');
+ const add = screen.getByTestId('mcp-consent-add') as HTMLButtonElement;
await waitFor(() => {
- expect(harness.confirmCalls).toHaveLength(1);
+ expect(add.disabled).toBe(true);
});
- });
-
- test('zero resolved hosts unchecks and disables the skill checkboxes with a reason', async () => {
- await renderDialog(makeHarness({ snapshot: skillsNoHostsPayload }));
- for (const id of ['discovery', 'write-skill']) {
- const box = screen.getByTestId(`mcp-consent-skill-checkbox-${id}`);
- expect(box.getAttribute('aria-checked')).toBe('false');
- expect(box.hasAttribute('disabled')).toBe(true);
- }
- // Each row states what would make it installable — the shared no-hosts copy.
- expect(screen.getAllByTestId('skill-consent-row-no-hosts')).toHaveLength(2);
- });
- test('a staged skill with zero hosts is never sent on Connect', async () => {
- const harness = await renderDialog(makeHarness({ snapshot: skillsNoHostsPayload }));
- // Skills were offered, so Add stays enabled; confirming sends an empty set
- // (nothing can land nowhere) rather than the pre-checked default.
- await userEvent.click(screen.getByTestId('mcp-consent-add'));
+ first.resolve({ ok: false, error: 'Could not write marker' });
await waitFor(() => {
- expect(harness.confirmCalls).toEqual([
- { editorIds: ['claude'], pathInstall: undefined, skills: [] },
- ]);
+ expect(add.disabled).toBe(false);
});
+ expect(harness.toastErrors).toEqual(['Could not write marker']);
});
});
diff --git a/packages/app/src/components/McpConsentDialog.test.ts b/packages/app/src/components/McpConsentDialog.test.ts
index 2bfd5802c..0e7fd031b 100644
--- a/packages/app/src/components/McpConsentDialog.test.ts
+++ b/packages/app/src/components/McpConsentDialog.test.ts
@@ -1,29 +1,13 @@
import { describe, expect, test, vi } from 'vitest';
-import type { OkMcpWiringEditorId, OkMcpWiringShowPayload } from '@/lib/desktop-bridge-types';
+import type { OkMcpWiringShowPayload } from '@/lib/desktop-bridge-types';
import { McpConsentDialog } from './McpConsentDialog';
-import {
- computeInitialSelection,
- computeInitialSkillSelection,
- isPathRowActionable,
- partitionEditorsForDisplay,
- selectedIdsOrdered,
- type ToastImpl,
- toggleSelectedId,
-} from './McpConsentDialogBody';
+import { connectableEditors, isPathRowActionable, type ToastImpl } from './McpConsentDialogBody';
type EditorDetection = OkMcpWiringShowPayload['detectedEditors'][number];
-type GlobalSkill = OkMcpWiringShowPayload['globalSkills'][number];
-
-/** Skill-descriptor factory — fills the disclosure fields the initial-selection
- * helper doesn't read (name/description/paths/size), so each case stays focused
- * on id + hosts. */
-function gs(o: Pick<GlobalSkill, 'id' | 'hosts'>): GlobalSkill {
- return { name: o.id, alreadyInstalled: false, description: '', paths: [], ...o };
-}
/** Detection-literal factory — fills the location-disclosure fields
- * (configPath/entryLocator) the pure selection/ordering helpers under test
- * don't read, so each case stays focused on id/detected/willReplace. */
+ * (configPath/entryLocator) the pure helpers under test don't read, so each
+ * case stays focused on id/detected/willReplace. */
function ed(o: Omit<EditorDetection, 'configPath' | 'entryLocator'>): EditorDetection {
return { configPath: null, entryLocator: 'mcpServers.open-knowledge', ...o };
}
@@ -35,117 +19,29 @@ const sampleDetection: readonly EditorDetection[] = [
ed({ id: 'codex', label: 'Codex', detected: false, willReplace: false }),
];
-describe('computeInitialSelection', () => {
- test('preselects every detected editor; skips undetected ones (OQ-14)', () => {
- const sel = computeInitialSelection(sampleDetection);
- expect(sel.has('claude')).toBe(true);
- expect(sel.has('cursor')).toBe(true);
- expect(sel.has('claude-desktop')).toBe(false);
- expect(sel.has('codex')).toBe(false);
- expect(sel.size).toBe(2);
- });
-
- test('empty payload yields empty selection', () => {
- const sel = computeInitialSelection([]);
- expect(sel.size).toBe(0);
+describe('connectableEditors', () => {
+ test('keeps detected tools in payload order and drops undetected ones', () => {
+ expect(connectableEditors(sampleDetection).map((e) => e.id)).toEqual(['claude', 'cursor']);
});
- test('all-detected preselects all', () => {
- const sel = computeInitialSelection([
- ed({ id: 'claude', label: 'Claude', detected: true, willReplace: false }),
- ed({ id: 'cursor', label: 'Cursor', detected: true, willReplace: false }),
- ]);
- expect(sel.size).toBe(2);
+ test('empty payload yields an empty write set', () => {
+ expect(connectableEditors([])).toEqual([]);
});
- test('none-detected preselects none', () => {
- const sel = computeInitialSelection([
+ test('none detected yields an empty write set — the dialog renders no checkbox', () => {
+ const out = connectableEditors([
ed({ id: 'claude', label: 'Claude', detected: false, willReplace: false }),
ed({ id: 'cursor', label: 'Cursor', detected: false, willReplace: false }),
]);
- expect(sel.size).toBe(0);
- });
-});
-
-describe('computeInitialSkillSelection', () => {
- test('pre-checks bundles with a resolved host; excludes a zero-host bundle', () => {
- const sel = computeInitialSkillSelection([
- gs({ id: 'discovery', hosts: ['claude'] }),
- gs({ id: 'write-skill', hosts: [] }),
- ]);
- expect(sel.has('discovery')).toBe(true);
- expect(sel.has('write-skill')).toBe(false);
- expect(sel.size).toBe(1);
- });
-
- test('all reach resolved: every bundle pre-checked (opt-out default)', () => {
- const sel = computeInitialSkillSelection([
- gs({ id: 'discovery', hosts: ['claude'] }),
- gs({ id: 'write-skill', hosts: ['cursor'] }),
- ]);
- expect(sel.size).toBe(2);
- });
-
- test('zero hosts everywhere: nothing pre-checked (never stage an install that cannot land)', () => {
- const sel = computeInitialSkillSelection([
- gs({ id: 'discovery', hosts: [] }),
- gs({ id: 'write-skill', hosts: [] }),
- ]);
- expect(sel.size).toBe(0);
- });
-});
-
-describe('toggleSelectedId', () => {
- test('adds id when absent', () => {
- const next = toggleSelectedId(new Set<OkMcpWiringEditorId>(), 'claude');
- expect(next.has('claude')).toBe(true);
- expect(next.size).toBe(1);
- });
-
- test('removes id when present', () => {
- const prev = new Set<OkMcpWiringEditorId>(['claude']);
- const next = toggleSelectedId(prev, 'claude');
- expect(next.has('claude')).toBe(false);
- expect(next.size).toBe(0);
- });
-
- test('returns a new Set (does not mutate input — immutable-style)', () => {
- const prev = new Set<OkMcpWiringEditorId>(['claude']);
- const next = toggleSelectedId(prev, 'cursor');
- expect(prev.has('cursor')).toBe(false);
- expect(next.has('cursor')).toBe(true);
- expect(prev).not.toBe(next);
- });
-});
-
-describe('selectedIdsOrdered', () => {
- test('projects selection back into array preserving detection order', () => {
- const sel = new Set<OkMcpWiringEditorId>(['cursor', 'claude']);
- const out = selectedIdsOrdered(sel, sampleDetection);
- // Detection order is [claude, claude-desktop, cursor, codex].
- // Projection keeps that order, dropping unselected entries.
- expect(out).toEqual(['claude', 'cursor']);
- });
-
- test('empty selection yields empty array', () => {
- const out = selectedIdsOrdered(new Set<OkMcpWiringEditorId>(), sampleDetection);
- expect(out).toEqual([]);
- });
-
- test('selected ids NOT in detection payload are dropped (defensive)', () => {
- const sel = new Set<OkMcpWiringEditorId>(['claude']);
- const truncated: readonly EditorDetection[] = [
- ed({ id: 'codex', label: 'Codex', detected: false, willReplace: false }),
- ];
- const out = selectedIdsOrdered(sel, truncated);
expect(out).toEqual([]);
});
- test('all-selected yields full detection order', () => {
- const allIds = sampleDetection.map((d) => d.id);
- const sel = new Set<OkMcpWiringEditorId>(allIds);
- const out = selectedIdsOrdered(sel, sampleDetection);
- expect(out).toEqual(allIds);
+ test('all detected yields every tool', () => {
+ const out = connectableEditors([
+ ed({ id: 'claude', label: 'Claude', detected: true, willReplace: false }),
+ ed({ id: 'cursor', label: 'Cursor', detected: true, willReplace: false }),
+ ]);
+ expect(out.map((e) => e.id)).toEqual(['claude', 'cursor']);
});
});
@@ -177,74 +73,11 @@ describe('isPathRowActionable', () => {
});
});
-describe('partitionEditorsForDisplay', () => {
- const claude = ed({ id: 'claude', label: 'Claude', detected: true, willReplace: false });
- const cursor = ed({ id: 'cursor', label: 'Cursor', detected: true, willReplace: false });
- const codex = ed({ id: 'codex', label: 'Codex', detected: false, willReplace: false });
- const opencode = ed({ id: 'opencode', label: 'OpenCode', detected: false, willReplace: false });
-
- const none = new Set<OkMcpWiringEditorId>();
-
- test('collapsed: shows only detected, hides unchecked undetected behind the toggle', () => {
- const r = partitionEditorsForDisplay([claude, codex, cursor, opencode], none, false);
- expect(r.collapsible).toBe(true);
- expect(r.hiddenCount).toBe(2);
- expect(r.visible.map((e) => e.id)).toEqual(['claude', 'cursor']);
- });
-
- test('expanded: always-shown first, then hideable; hiddenCount labels the toggle', () => {
- const r = partitionEditorsForDisplay([claude, codex, cursor, opencode], none, true);
- expect(r.collapsible).toBe(true);
- expect(r.hiddenCount).toBe(2);
- expect(r.visible.map((e) => e.id)).toEqual(['claude', 'cursor', 'codex', 'opencode']);
- });
-
- test('consent integrity: a checked undetected tool survives collapse (never hidden)', () => {
- // User expanded, checked codex (undetected), then collapsed. codex must stay
- // visible because it is still in the write set; only unchecked opencode hides.
- const r = partitionEditorsForDisplay(
- [claude, codex, cursor, opencode],
- new Set<OkMcpWiringEditorId>(['codex']),
- false,
- );
- expect(r.collapsible).toBe(true);
- expect(r.hiddenCount).toBe(1);
- expect(r.visible.map((e) => e.id)).toEqual(['claude', 'codex', 'cursor']);
- });
-
- test('all detected: not collapsible, shows everything', () => {
- const r = partitionEditorsForDisplay([claude, cursor], none, false);
- expect(r.collapsible).toBe(false);
- expect(r.hiddenCount).toBe(0);
- expect(r.visible.map((e) => e.id)).toEqual(['claude', 'cursor']);
- });
-
- test('every undetected tool checked: nothing hideable, not collapsible', () => {
- const r = partitionEditorsForDisplay(
- [codex, opencode],
- new Set<OkMcpWiringEditorId>(['codex', 'opencode']),
- false,
- );
- expect(r.collapsible).toBe(false);
- expect(r.hiddenCount).toBe(0);
- expect(r.visible.map((e) => e.id)).toEqual(['codex', 'opencode']);
- });
-
- test('none detected + none checked: empty-state fallback shows all (never a blank list)', () => {
- const r = partitionEditorsForDisplay([codex, opencode], none, false);
- expect(r.collapsible).toBe(false);
- expect(r.hiddenCount).toBe(0);
- expect(r.visible.map((e) => e.id)).toEqual(['codex', 'opencode']);
- });
-});
-
describe('McpConsentDialog module shape', () => {
- test('exports the component + the pure selection/display helpers + ToastImpl type', () => {
+ test('exports the component + the pure write-set helpers + ToastImpl type', () => {
expect(typeof McpConsentDialog).toBe('function');
- expect(typeof computeInitialSelection).toBe('function');
- expect(typeof toggleSelectedId).toBe('function');
- expect(typeof selectedIdsOrdered).toBe('function');
- expect(typeof partitionEditorsForDisplay).toBe('function');
+ expect(typeof connectableEditors).toBe('function');
+ expect(typeof isPathRowActionable).toBe('function');
// ToastImpl is a type; no runtime export — this assertion just ensures
// the import resolves at type-check time. The shape is exercised by the
// toast injection contract below.
diff --git a/packages/app/src/components/McpConsentDialogBody.tsx b/packages/app/src/components/McpConsentDialogBody.tsx
index 6bb3022b0..2ade2d3cd 100644
--- a/packages/app/src/components/McpConsentDialogBody.tsx
+++ b/packages/app/src/components/McpConsentDialogBody.tsx
@@ -3,31 +3,43 @@
* so that file can lazy-load this module via `React.lazy()`. See that file's
* header for the why.
*
- * Minimum-viable UI: title, scrollable checkbox list of detected
- * editors (preselected — true if detection.detected), Add primary +
- * Skip secondary. ESC / outside-click = skip via shadcn Dialog's built-in
- * behavior (routed through `onOpenChange(false)` → skip()).
+ * One decision per independent consequence, and no more. The AI-tool setup is a
+ * SINGLE pre-checked checkbox covering every detected tool plus the user-global
+ * discovery skill, because those are one thought ("let my agents see this") and
+ * splitting them made a first-run screen the user had to audit row by row.
+ * Per-tool granularity lives in Settings → AI tools & CLI, where it can be
+ * revisited; here it would only slow down the common answer.
*
- * The dialog also gates the shell-PATH install: a distinct pre-checked
- * toggle in its own "Terminal" section, rendered first in the scrollable
- * body above the editor list, driven by `payload.pathInstall`. Hidden when no rc file is
- * touchable; informational when the managed block is already on disk /
- * consent already granted. Unchecking degrades only `ok` in EXTERNAL
- * terminals — OpenKnowledge's built-in terminal injects `~/.ok/bin` itself
- * and MCP wiring runs over npx, so the warning copy is scoped to exactly
- * that.
+ * The shell-PATH install stays its own checkbox: MCP runs over npx / the bundle
+ * wrapper and never over bare `ok`, so wanting one says nothing about wanting
+ * the other.
+ *
+ * Consent integrity — this dialog fires once, so it must disclose exactly what
+ * it will touch:
+ * - The checkbox label NAMES every tool in the write set, collapsed or not.
+ * - A replacement warning sits next to the checkbox, never behind the
+ * expander, so an overwrite is never something the user had to go looking
+ * for.
+ * - "What this changes" lists the exact files, entries and skill destinations.
+ * Every path comes from main's descriptors, which are computed from the
+ * installer's own iteration set and gates — never re-derived here.
+ *
+ * Undetected tools are absent entirely: a row that writes nothing is noise on a
+ * first-run screen, and Settings lists them all.
+ *
+ * The screen only ever ADDS. Leaving the box unchecked records no skill decision
+ * (`skills: undefined`, which main reads as "no decision" rather than "decline
+ * all"), so dismissing setup can never tear down a bundle already on disk.
+ * Removal is Settings' job, where the row states what it removes.
*/
// biome-ignore-all lint/plugin/no-physical-direction-utility: pre-rule backlog — physical margin/padding/inset utilities predate the rule; drain by swapping ml/mr → ms/me, pl/pr → ps/pe, left/right → start/end, then deleting this line. See https://github.com/inkeep/open-knowledge/blob/main/biome-plugins/README.md#no-physical-direction-utilitygrit
-import { EDITOR_SETUP_DOC_SLUG } from '@inkeep/open-knowledge-core';
-import { Plural, Trans, useLingui } from '@lingui/react/macro';
-import { ArrowUpRight, ChevronDown, ChevronUp, Info } from 'lucide-react';
+import { i18n } from '@lingui/core';
+import { Trans, useLingui } from '@lingui/react/macro';
+import { ChevronDown, ChevronUp, Info } from 'lucide-react';
import { useId, useState } from 'react';
import { toast as sonnerToast } from 'sonner';
-import { SkillConsentRow } from '@/components/SkillConsentRow';
-import { SkillCostValue } from '@/components/SkillCostValue';
-import { SkillDestinationList } from '@/components/SkillDestinationList';
import { Button } from '@/components/ui/button';
import { Checkbox } from '@/components/ui/checkbox';
import {
@@ -41,13 +53,12 @@ import {
} from '@/components/ui/dialog';
import { Label } from '@/components/ui/label';
import { Tooltip, TooltipContent, TooltipTrigger } from '@/components/ui/tooltip';
-import type { OkMcpWiringEditorId, OkMcpWiringShowPayload } from '@/lib/desktop-bridge-types';
-import { dispatchExternalLinkClick } from '@/lib/external-link';
+import type { OkMcpWiringShowPayload } from '@/lib/desktop-bridge-types';
import { type McpConsentStore, mcpConsentStore } from '@/lib/mcp-consent-store';
+import { formatToolList } from '@/lib/tool-list-format';
type EditorDetection = OkMcpWiringShowPayload['detectedEditors'][number];
type PathInstallDescriptor = OkMcpWiringShowPayload['pathInstall'];
-type GlobalSkillDescriptor = OkMcpWiringShowPayload['globalSkills'][number];
/**
* Pure helper: whether the PATH row solicits a decision. Hidden rows
@@ -60,102 +71,14 @@ export function isPathRowActionable(pathInstall: PathInstallDescriptor): boolean
}
/**
- * Pure helper: from the detection payload, compute the initial checkbox
- * state — each detected editor starts checked, undetected
- * editors start unchecked but still appear in the list.
- */
-export function computeInitialSelection(
- detectedEditors: readonly EditorDetection[],
-): ReadonlySet<OkMcpWiringEditorId> {
- const out = new Set<OkMcpWiringEditorId>();
- for (const d of detectedEditors) if (d.detected) out.add(d.id);
- return out;
-}
-
-/** Pure helper: toggle a checkbox; returns a new Set (immutable-style). */
-export function toggleSelectedId(
- prev: ReadonlySet<OkMcpWiringEditorId>,
- id: OkMcpWiringEditorId,
-): ReadonlySet<OkMcpWiringEditorId> {
- const next = new Set(prev);
- if (next.has(id)) {
- next.delete(id);
- } else {
- next.add(id);
- }
- return next;
-}
-
-/**
- * Pure helper: project the selected Set back into an array preserving the
- * detection payload's order. Used at confirm time so downstream writes iterate
- * editors in the same order the user saw them.
- */
-export function selectedIdsOrdered(
- selection: ReadonlySet<OkMcpWiringEditorId>,
- detectedEditors: readonly EditorDetection[],
-): OkMcpWiringEditorId[] {
- const out: OkMcpWiringEditorId[] = [];
- for (const d of detectedEditors) if (selection.has(d.id)) out.push(d.id);
- return out;
-}
-
-/**
- * Pure helper: progressive-disclosure split for the editor list. Rows that are
- * detected OR currently checked always show; the rest hide behind a "Show N
- * more" toggle until `showAll`. Keying visibility off `selection` (not just
- * `detected`) is load-bearing for consent integrity: a checked undetected tool
- * must never be collapsed out of view while it's still in the write set — this
- * dialog discloses exactly what Connect will touch, and it fires once. So
- * `hiddenCount` is the count of rows actually hidden (unchecked + undetected),
- * which keeps the toggle label truthful. `collapsible` is false when the split
- * is pointless — nothing hideable, or nothing always-shown (which would blank
- * the list) — in which case every row shows.
+ * Pure helper: the write set. Detected tools only, in payload order — an
+ * undetected tool has no config to wire, so including it would name a
+ * destination nothing is written to.
*/
-export function partitionEditorsForDisplay(
+export function connectableEditors(
editors: readonly EditorDetection[],
- selection: ReadonlySet<OkMcpWiringEditorId>,
- showAll: boolean,
-): { visible: readonly EditorDetection[]; hiddenCount: number; collapsible: boolean } {
- const alwaysShown = editors.filter((e) => e.detected || selection.has(e.id));
- const hideable = editors.filter((e) => !e.detected && !selection.has(e.id));
- if (alwaysShown.length === 0 || hideable.length === 0) {
- return { visible: editors, hiddenCount: 0, collapsible: false };
- }
- return {
- visible: showAll ? [...alwaysShown, ...hideable] : alwaysShown,
- hiddenCount: hideable.length,
- collapsible: true,
- };
-}
-
-/**
- * Pure helper: initial skill checkbox state — every offered bundle with a place
- * to land starts checked (opt-out default: preserves today's install-everywhere
- * behavior while making it one-click-off). A bundle whose reach resolved to zero
- * hosts starts unchecked; its row disables the checkbox, so it can never be
- * staged for an install that would land nowhere.
- */
-export function computeInitialSkillSelection(
- globalSkills: readonly GlobalSkillDescriptor[],
-): ReadonlySet<string> {
- return new Set(globalSkills.filter((s) => s.hosts.length > 0).map((s) => s.id));
-}
-
-/** Pure helper: toggle a skill checkbox; returns a new Set. */
-export function toggleSkillId(prev: ReadonlySet<string>, id: string): ReadonlySet<string> {
- const next = new Set(prev);
- if (next.has(id)) next.delete(id);
- else next.add(id);
- return next;
-}
-
-/** Pure helper: project the checked skills back into payload order. */
-export function skillIdsOrdered(
- selection: ReadonlySet<string>,
- globalSkills: readonly GlobalSkillDescriptor[],
-): string[] {
- return globalSkills.filter((s) => selection.has(s.id)).map((s) => s.id);
+): readonly EditorDetection[] {
+ return editors.filter((e) => e.detected);
}
/**
@@ -206,53 +129,6 @@ export function McpConsentDialogBody({
return <McpConsentDialogForm payload={snapshot} store={store} toast={toast} />;
}
-/**
- * Per-editor location disclosure — the exact config file + entry an Add would
- * touch, shown behind an info affordance. Mirrors Settings → AI tools'
- * `RowInfoTooltip` (reuses the same "File" / "Entry" copy) so the first-launch
- * dialog and the persistent settings surface disclose identically — including
- * relying on the root `TooltipProvider` (main.tsx) rather than wrapping a
- * per-row one, so sibling info buttons share Radix's skip-delay window.
- */
-function EditorLocationTooltip({ editor }: { editor: EditorDetection }) {
- const { t } = useLingui();
- return (
- <Tooltip>
- <TooltipTrigger asChild>
- <Button
- variant="ghost"
- size="icon"
- className="mt-1.5 mr-1.5 size-6 shrink-0 text-muted-foreground opacity-60 hover:opacity-100"
- aria-label={t`What this checkbox changes`}
- data-testid={`mcp-consent-editor-info-${editor.id}`}
- >
- <Info className="size-3.5" aria-hidden />
- </Button>
- </TooltipTrigger>
- {/* Base TooltipContent is an inline-flex row — the column wrapper keeps
- the File/Entry pairs stacked instead of side by side. */}
- <TooltipContent side="left" className="max-w-sm text-left">
- <div className="flex min-w-0 flex-col gap-1">
- <p className="opacity-70">
- <Trans>File</Trans>
- </p>
- <p>
- <code className="break-all">
- {editor.configPath ?? t`unavailable on this platform`}
- </code>
- </p>
- <p className="pt-1 opacity-70">
- <Trans>Entry</Trans>
- </p>
- <p>
- <code className="break-all">{editor.entryLocator}</code>
- </p>
- </div>
- </TooltipContent>
- </Tooltip>
- );
-}
-
interface McpConsentDialogFormProps {
payload: OkMcpWiringShowPayload;
store: McpConsentStore;
@@ -261,64 +137,49 @@ interface McpConsentDialogFormProps {
function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormProps) {
const { t } = useLingui();
- const detectedEditors = payload.detectedEditors;
const pathInstall = payload.pathInstall;
const globalSkills = payload.globalSkills;
const skillsOffered = globalSkills.length > 0;
const pathActionable = isPathRowActionable(pathInstall);
- const [selection, setSelection] = useState<ReadonlySet<OkMcpWiringEditorId>>(() =>
- computeInitialSelection(detectedEditors),
- );
+ const editors = connectableEditors(payload.detectedEditors);
+ const hasEditors = editors.length > 0;
+ const replacing = editors.filter((e) => e.willReplace);
+ // Pre-checked (opt-out): the common answer is yes, and the label names
+ // everything it covers so agreeing isn't agreeing blind.
+ const [connectChecked, setConnectChecked] = useState(true);
// Pre-checked (opt-out) when the row solicits a decision; informational
// rows render force-checked + disabled below and never read this state.
const [pathChecked, setPathChecked] = useState(true);
- // Pre-checked (opt-out) — every offered bundle with a resolved host starts on.
- const [skillSelection, setSkillSelection] = useState<ReadonlySet<string>>(() =>
- computeInitialSkillSelection(globalSkills),
- );
- // Which skill rows are expanded to their full cost + destination disclosure.
- // Purely presentational — expansion never touches the staged selection.
- const [expandedSkills, setExpandedSkills] = useState<ReadonlySet<string>>(() => new Set());
+ const [showDetails, setShowDetails] = useState(false);
const [busy, setBusy] = useState(false);
- // Progressive disclosure: detected tools show by default; undetected ones hide
- // behind a "Show N more" toggle (see partitionEditorsForDisplay for the
- // empty-state fallback that shows all when nothing is detected).
- const [showAllEditors, setShowAllEditors] = useState(false);
- const editorList = partitionEditorsForDisplay(detectedEditors, selection, showAllEditors);
const idPrefix = useId();
- function onToggle(id: OkMcpWiringEditorId) {
- setSelection((prev) => toggleSelectedId(prev, id));
- }
-
- function toggleSkillExpanded(id: string) {
- setExpandedSkills((prev) => {
- const next = new Set(prev);
- if (next.has(id)) next.delete(id);
- else next.add(id);
- return next;
- });
- }
-
- async function onAdd() {
+ async function onContinue() {
setBusy(true);
+ const connecting = connectChecked && hasEditors;
const result = await store.confirm({
- editorIds: selectedIdsOrdered(selection, detectedEditors),
+ editorIds: connecting ? editors.map((e) => e.id) : [],
pathInstall: pathActionable ? pathChecked : undefined,
- // When skills are offered, always send the (possibly empty) selection so
- // main records a decision for every bundle — an empty list declines both.
- skills: skillsOffered ? skillIdsOrdered(skillSelection, globalSkills) : undefined,
+ // An array records a decision for every offered bundle; `undefined` records
+ // none. Declining setup must leave an existing install alone, so the
+ // decline path sends `undefined`, never `[]`.
+ skills: connecting && skillsOffered ? globalSkills.map((s) => s.id) : undefined,
});
// Success: the store clears `currentRequest` → useSyncExternalStore
// unmounts this subtree, so there's nothing to reset. Failure
// (ok:false / thrown rejection): the store KEEPS the snapshot
- // populated, so we must reset
- // `busy` here or the Add button stays disabled forever and same-boot
- // retry is impossible. Sonner is mounted globally in main.tsx; the
- // toast surfaces even if the dialog were to unmount.
+ // populated, so we must reset `busy` here or the button stays disabled
+ // forever and same-boot retry is impossible. Sonner is mounted globally
+ // in main.tsx; the toast surfaces even if the dialog were to unmount.
if (!result.ok) {
toast.error(result.error);
setBusy(false);
+ return;
+ }
+ // Continuing without connecting anything is a legitimate choice, but the
+ // dialog is one-shot — without this the surface is easy to lose track of.
+ if (!connecting) {
+ toast.message(t`This can be configured in Settings > AI tools & CLI`);
}
}
@@ -326,20 +187,18 @@ function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormPro
setBusy(true);
const result = await store.skip();
if (result.ok) {
- // Point the user at where these same choices live for later — the
- // dialog is one-shot per boot, so without this the surface is easy
- // to lose track of.
toast.message(t`This can be configured in Settings > AI tools & CLI`);
} else {
toast.error(result.error);
- // Matching rationale to onAdd — reset `busy` so Skip stays
- // clickable after a transient marker-write failure.
+ // Matching rationale to onContinue — reset `busy` so the dialog stays
+ // usable after a transient marker-write failure.
setBusy(false);
}
}
function onOpenChange(open: boolean) {
- // ESC, outside-click, X button — treat as skip.
+ // ESC, outside-click, X button — no decision was made, so this is a skip
+ // (marker only), not a decline.
if (!open && !busy) void onSkip();
}
@@ -349,21 +208,21 @@ function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormPro
* Radix Dialog auto-wires `aria-labelledby` / `aria-describedby` on
* `DialogContent` from `DialogTitle` / `DialogDescription` via context
* — no manual `useId` plumbing needed. Each row's `<Label>` is
- * associated to its `<Checkbox>` by
- * `htmlFor` + matching `id`, providing the accessible name; no
- * `aria-describedby` on the checkbox itself, since duplicating the
- * label content via that attr causes screen readers to either
- * announce the label twice or drop the association.
+ * associated to its `<Checkbox>` by `htmlFor` + matching `id`,
+ * providing the accessible name; no `aria-describedby` on the checkbox
+ * itself, since duplicating the label content via that attr causes
+ * screen readers to either announce the label twice or drop the
+ * association.
*/}
- <DialogContent className="sm:max-w-lg">
+ <DialogContent className="sm:max-w-lg" aria-busy={busy}>
<DialogHeader>
<DialogTitle>
<Trans>Connect your AI tools to OpenKnowledge</Trans>
</DialogTitle>
<DialogDescription>
<Trans>
- Give the AI tools you use access to read and update your projects. Pick what to set up
- below, and change it anytime in Settings &gt; AI tools & CLI.
+ Give the AI tools you use access to read and update your projects. You can change any
+ of this later in Settings &gt; AI tools & CLI.
</Trans>
</DialogDescription>
</DialogHeader>
@@ -371,11 +230,11 @@ function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormPro
<DialogBody className="flex flex-col gap-6 min-h-0">
{/*
* Shell-PATH consent section — rendered first inside the scrollable
- * DialogBody, above the editor list. Distinct from the per-editor
- * MCP checkboxes because the two decisions are independent (MCP runs
- * over npx / the bundle wrapper, never bare `ok`). Hidden when no rc
- * file is touchable; informational when a managed block is already
- * on disk or consent was already granted.
+ * DialogBody, above the AI-tools row. Distinct from the AI-tools
+ * checkbox because the two decisions are independent (MCP runs over
+ * npx / the bundle wrapper, never bare `ok`). Hidden when no rc file
+ * is touchable; informational when a managed block is already on disk
+ * or consent was already granted.
*/}
{pathInstall.shellDetected && (
<div className="flex flex-col gap-1.5">
@@ -413,12 +272,12 @@ function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormPro
</Trans>
{pathActionable && (
<Tooltip>
- {/* Relies on the root TooltipProvider (main.tsx), same as
- EditorLocationTooltip — so every info button in this dialog
- shares one skip-delay window. Nested inside the row <Label>,
- so stop the click from bubbling to toggle the checkbox;
- Radix opens on hover/focus (not click), so preventing the
- click default costs nothing. */}
+ {/* Relies on the root TooltipProvider (main.tsx) so every
+ info button in this dialog shares one skip-delay window.
+ Nested inside the row <Label>, so stop the click from
+ bubbling to toggle the checkbox; Radix opens on
+ hover/focus (not click), so preventing the click default
+ costs nothing. */}
<TooltipTrigger
onClick={(e) => {
e.preventDefault();
@@ -452,7 +311,7 @@ function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormPro
)}
{pathActionable && !pathChecked && (
<span
- className="ps-6.5 text-xs text-amber-600 dark:text-amber-400"
+ className="ps-6.5 text-xs text-amber-700 dark:text-amber-400"
data-testid="mcp-consent-path-warning"
>
<Trans comment="Warning shown when the user unchecks the PATH toggle — only external terminals degrade">
@@ -466,238 +325,196 @@ function McpConsentDialogForm({ payload, store, toast }: McpConsentDialogFormPro
</div>
</div>
)}
+
+ {/*
+ * The AI-tools decision. One checkbox for every detected tool plus the
+ * offered skill bundles; the label names the tools so the collapsed
+ * state still discloses the write set.
+ */}
<div className="flex flex-col gap-1.5">
- {/* Group label only when the Terminal section renders above —
- with a single group there is nothing to distinguish. */}
{pathInstall.shellDetected && (
<div className="text-xs font-medium text-muted-foreground">
- <Trans comment="Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool">
- MCP connections
+ <Trans comment="Section label above the AI-tools checkbox in the first-launch dialog">
+ AI tools
</Trans>
</div>
)}
- <ul className="rounded-md border border-border bg-card/50 divide-y divide-border overflow-hidden">
- {editorList.visible.map((editor) => {
- const checked = selection.has(editor.id);
- const checkboxId = `${idPrefix}-${editor.id}`;
- const setupUrl = `https://openknowledge.ai/docs/integrations/${EDITOR_SETUP_DOC_SLUG[editor.id]}`;
- return (
- // The <Label> owns the full row (flex-1) and height (py-2.5) so the
- // whole name area toggles the checkbox. The setup link and location
- // info button are siblings OUTSIDE the label (an anchor / a second
- // interactive control must not live in the checkbox's activation
- // path).
- <li key={editor.id} className="flex items-start hover:bg-accent">
- <Label
- htmlFor={checkboxId}
- className="flex flex-1 cursor-pointer items-start gap-2.5 px-3 py-2.5 font-normal"
+ {hasEditors ? (
+ <div className="overflow-hidden rounded-md border border-border bg-card/50">
+ <Label
+ htmlFor={`${idPrefix}-connect`}
+ className="flex cursor-pointer items-start gap-2.5 px-3 py-2.5 font-normal hover:bg-accent"
+ >
+ <Checkbox
+ id={`${idPrefix}-connect`}
+ checked={connectChecked}
+ disabled={busy}
+ onCheckedChange={() => setConnectChecked((prev) => !prev)}
+ className="mt-0.5"
+ data-testid="mcp-consent-connect-checkbox"
+ />
+ <span className="flex min-w-0 flex-1 flex-col gap-0.5">
+ <span
+ className="text-sm font-medium text-foreground"
+ data-testid="mcp-consent-connect-summary"
>
- <Checkbox
- id={checkboxId}
- checked={checked}
- disabled={busy}
- onCheckedChange={() => onToggle(editor.id)}
- className="mt-0.5"
- data-testid={`mcp-consent-checkbox-${editor.id}`}
- />
- <span className="flex min-w-0 flex-1 flex-col gap-0.5">
- <span className="text-sm font-medium text-foreground">{editor.label}</span>
- {/* willReplace warns Add will overwrite an existing OK-managed
- entry (reclaimed by name). Sits under the name, inside the
- label column, so it aligns with the Settings status subtext. */}
- {editor.willReplace && (
- <span
- className="text-xs text-amber-600 dark:text-amber-400"
- data-testid={`mcp-consent-status-${editor.id}`}
- >
- <Trans comment="Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry">
- Will replace existing OpenKnowledge entry
- </Trans>
- </span>
- )}
- </span>
- </Label>
- {/* Undetected tools link to their setup guide instead of a
- dead-end "Not detected"; detected tools need no trailing line. */}
- {!editor.willReplace && !editor.detected && (
- <a
- href={setupUrl}
- target="_blank"
- rel="noopener noreferrer"
- onClick={(e) => dispatchExternalLinkClick(e, setupUrl)}
- onAuxClick={(e) => dispatchExternalLinkClick(e, setupUrl)}
- // Per-tool name so a screen-reader link list distinguishes rows
- // (2.4.4); contains the visible text (2.5.3) and flags the
- // new-tab behavior the arrow icon shows sighted users.
- aria-label={t`How to set up ${editor.label} (opens in browser)`}
- className="flex shrink-0 items-center gap-0.5 px-2 py-2.5 text-xs text-muted-foreground underline-offset-2 hover:text-foreground hover:underline"
- data-testid={`mcp-consent-status-${editor.id}`}
+ {t`Set up ${formatToolList(
+ editors.map((e) => e.label),
+ i18n.locale,
+ )}`}
+ </span>
+ <span className="text-xs text-muted-foreground">
+ {skillsOffered ? (
+ <Trans comment="Subtext under the single AI-tools checkbox when a skill is also installed">
+ Adds OpenKnowledge's MCP connection to each, plus the
+ open-knowledge-discovery skill so your agents recognize OpenKnowledge
+ projects.
+ </Trans>
+ ) : (
+ <Trans comment="Subtext under the single AI-tools checkbox when only MCP is wired">
+ Adds OpenKnowledge's MCP connection to each.
+ </Trans>
+ )}
+ </span>
+ {/* Overwrite disclosure lives beside the checkbox, never behind
+ the expander — replacing a config the user never saw named is
+ the one outcome a collapsed summary must not hide. Gated on
+ the checkbox too: "Replaces …" is present tense, so leaving it
+ up after an uncheck states a consequence that will not happen,
+ and reads to a consent-conscious user as the uncheck not
+ having taken. The expander stays ungated — it describes what
+ the setup writes, which is the thing being decided about. */}
+ {connectChecked && replacing.length > 0 && (
+ <span
+ className="text-xs text-amber-700 dark:text-amber-400"
+ data-testid="mcp-consent-connect-replace-warning"
>
- <Trans comment="Link on an undetected tool row to its OpenKnowledge setup guide">
- How to set up
+ {t`Replaces the existing OpenKnowledge entry in ${formatToolList(
+ replacing.map((e) => e.label),
+ i18n.locale,
+ )}`}
+ </span>
+ )}
+ </span>
+ </Label>
+ <div className="border-t border-border">
+ <Button
+ type="button"
+ variant="ghost"
+ size="sm"
+ className="h-auto w-full justify-start gap-1 rounded-none px-3 py-2 text-xs font-normal text-muted-foreground hover:text-foreground"
+ onClick={() => setShowDetails((v) => !v)}
+ data-testid="mcp-consent-details-toggle"
+ aria-expanded={showDetails}
+ aria-controls={`${idPrefix}-details`}
+ >
+ {showDetails ? (
+ <>
+ <ChevronUp className="size-3.5" aria-hidden />
+ <Trans comment="Collapses the list of files the setup writes to">
+ Hide details
+ </Trans>
+ </>
+ ) : (
+ <>
+ <ChevronDown className="size-3.5" aria-hidden />
+ <Trans comment="Expands the list of files the setup writes to">
+ What this changes
</Trans>
- <ArrowUpRight className="size-3" aria-hidden />
- </a>
+ </>
)}
- <EditorLocationTooltip editor={editor} />
- </li>
- );
- })}
- </ul>
- {/* Progressive disclosure toggle — only when the list has both
- detected and undetected tools, so there is something to hide/reveal. */}
- {editorList.collapsible && (
- <Button
- type="button"
- variant="ghost"
- size="sm"
- className="h-auto justify-start gap-1 self-start px-1 py-1 text-xs font-normal text-muted-foreground hover:text-foreground"
- onClick={() => setShowAllEditors((v) => !v)}
- data-testid="mcp-consent-editors-toggle"
- aria-expanded={showAllEditors}
- >
- {showAllEditors ? (
- <>
- <ChevronUp className="size-3.5" aria-hidden />
- <Trans comment="Collapses the undetected AI tools back behind the toggle">
- Show fewer
- </Trans>
- </>
- ) : (
- <>
- <ChevronDown className="size-3.5" aria-hidden />
- <Plural
- value={editorList.hiddenCount}
- one="Show # more tool"
- other="Show # more tools"
- />
- </>
- )}
- </Button>
- )}
- </div>
- {/*
- * User-global Agent Skills consent section — one staged, pre-checked
- * row per bundle rendered through the shared SkillConsentRow, so this
- * surface and Settings disclose a skill identically (its own
- * description, reach cluster, and cost). Distinct from the editor list
- * because skills install to every detected host by design, not per
- * editor — the fan-out note states that independence. Nothing writes on
- * check; the single Connect commits the whole staged set. Clicking a
- * row body expands it in place — never leaving this modal — to the full
- * three-tier cost and destination list the Settings confirm modal
- * carries. A bundle whose reach resolved to zero hosts renders its
- * checkbox unchecked + disabled with the row's own no-hosts copy, so
- * first launch never stages an install that cannot land.
- */}
- {skillsOffered && (
- <div className="flex flex-col gap-1.5">
- <span className="text-xs font-medium text-muted-foreground">
- <Trans comment="Section label above the skill checkboxes in the first-launch dialog">
- Agent Skills
- </Trans>
- </span>
+ </Button>
+ {/* The region element is always present so the toggle's
+ `aria-controls` never dangles — axe flags a reference to an
+ id that isn't in the DOM, which is exactly what a
+ conditionally-mounted target produces. */}
+ <div id={`${idPrefix}-details`}>
+ {showDetails && (
+ <div
+ className="flex flex-col gap-3 border-t border-border px-3 py-2.5"
+ data-testid="mcp-consent-details"
+ >
+ <div className="flex flex-col gap-1">
+ <span className="text-xs font-medium text-muted-foreground">
+ <Trans comment="Heading above the per-tool MCP config paths in the disclosure">
+ MCP connections
+ </Trans>
+ </span>
+ <ul className="flex flex-col gap-1.5">
+ {editors.map((editor) => (
+ <li
+ key={editor.id}
+ className="flex min-w-0 flex-col"
+ data-testid={`mcp-consent-detail-${editor.id}`}
+ >
+ <span className="text-1sm text-foreground">{editor.label}</span>
+ <span className="text-xs text-muted-foreground wrap-break-word">
+ <code className="break-all">
+ {editor.configPath ?? t`unavailable on this platform`}
+ </code>{' '}
+ · <code className="break-all">{editor.entryLocator}</code>
+ </span>
+ </li>
+ ))}
+ </ul>
+ </div>
+ {globalSkills.map((skill) => (
+ <div
+ key={skill.id}
+ className="flex flex-col gap-1"
+ data-testid={`mcp-consent-detail-skill-${skill.id}`}
+ >
+ <span className="text-xs font-medium text-muted-foreground">
+ <Trans comment="Heading above the skill install destinations in the disclosure">
+ Agent Skill
+ </Trans>
+ </span>
+ <span className="text-1sm text-foreground">
+ <code>{skill.name}</code>
+ </span>
+ <ul className="flex flex-col">
+ {skill.paths.map((path) => (
+ <li
+ key={path}
+ className="text-xs text-muted-foreground wrap-break-word"
+ >
+ <code className="break-all">{path}</code>
+ </li>
+ ))}
+ </ul>
+ </div>
+ ))}
+ </div>
+ )}
+ </div>
+ </div>
+ </div>
+ ) : (
+ /* Nothing detected: no checkbox, because there is nothing to write.
+ The Terminal row above may still be actionable, so the dialog
+ itself stays. */
<p
- className="text-xs text-muted-foreground"
- data-testid="mcp-consent-skill-fanout-note"
+ className="rounded-md border border-border bg-card/50 px-3 py-2.5 text-1sm text-muted-foreground"
+ data-testid="mcp-consent-no-tools"
>
- <Trans comment="Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above">
- Skills install to every AI tool detected on this machine, independent of the MCP
- connections you chose above.
+ <Trans comment="Shown in place of the AI-tools checkbox when no AI tool was detected">
+ No AI tools detected yet. Once you install one, connect it from Settings &gt; AI
+ tools & CLI.
</Trans>
</p>
- <ul className="rounded-md border border-border bg-card/50 divide-y divide-border overflow-hidden">
- {globalSkills.map((skill) => {
- const checked = skillSelection.has(skill.id);
- const hasHosts = skill.hosts.length > 0;
- const expanded = expandedSkills.has(skill.id);
- const checkboxId = `${idPrefix}-skill-${skill.id}`;
- return (
- <li key={skill.id} className="hover:bg-accent">
- <SkillConsentRow
- name={skill.name}
- description={skill.description}
- hosts={skill.hosts}
- size={skill.size}
- onActivate={() => toggleSkillExpanded(skill.id)}
- ariaExpanded={expanded}
- control={
- <Checkbox
- id={checkboxId}
- checked={hasHosts && checked}
- disabled={busy || !hasHosts}
- onCheckedChange={() =>
- setSkillSelection((prev) => toggleSkillId(prev, skill.id))
- }
- aria-label={t`Set up ${skill.name}`}
- data-testid={`mcp-consent-skill-checkbox-${skill.id}`}
- />
- }
- />
- {skill.alreadyInstalled && hasHosts && !checked && (
- <p
- role="status"
- className="px-3 pb-2.5 text-xs text-amber-600 dark:text-amber-400"
- data-testid={`mcp-consent-skill-warning-${skill.id}`}
- >
- <Trans comment="Warning shown when the user unchecks an already-installed skill">
- Removes this skill from your editors.
- </Trans>
- </p>
- )}
- {expanded && (
- <div
- className="flex flex-col gap-3 px-3 pb-3"
- data-testid={`mcp-consent-skill-expansion-${skill.id}`}
- >
- {skill.size ? (
- <div className="flex flex-col gap-1">
- <span className="text-xs font-medium text-muted-foreground">
- <Trans>Context cost</Trans>
- </span>
- <SkillCostValue size={skill.size} />
- </div>
- ) : null}
- {skill.paths.length > 0 && (
- <div className="flex flex-col gap-1">
- <span className="text-xs font-medium text-muted-foreground">
- <Trans>Installs to</Trans>
- </span>
- <SkillDestinationList paths={skill.paths} />
- </div>
- )}
- </div>
- )}
- </li>
- );
- })}
- </ul>
- </div>
- )}
+ )}
+ </div>
</DialogBody>
<DialogFooter>
- <Button
- variant="outline-mono"
- onClick={() => void onSkip()}
- disabled={busy}
- data-testid="mcp-consent-skip"
- >
- <Trans comment="Secondary button — dismisses the dialog without wiring any tools">
- Skip for now
- </Trans>
- </Button>
- <Button
- onClick={() => void onAdd()}
- disabled={
- busy || (selection.size === 0 && !(pathActionable && pathChecked) && !skillsOffered)
- }
- data-testid="mcp-consent-add"
- >
+ {/* One button: with a single checkbox, a separate "Skip" would be a
+ second way to say what unchecking already says. ESC / outside-click
+ still route to skip() — that path made no decision at all. */}
+ <Button onClick={() => void onContinue()} disabled={busy} data-testid="mcp-consent-add">
{busy ? (
<Trans>Working</Trans>
) : (
- <Trans comment="Primary button that writes MCP config for the selected AI tools">
- Connect
+ <Trans comment="Primary button that applies the first-launch setup choices">
+ Continue
</Trans>
)}
</Button>
diff --git a/packages/app/src/lib/tool-list-format.test.ts b/packages/app/src/lib/tool-list-format.test.ts
new file mode 100644
index 000000000..cb492d2f4
--- /dev/null
+++ b/packages/app/src/lib/tool-list-format.test.ts
@@ -0,0 +1,30 @@
+import { describe, expect, test } from 'vitest';
+import { formatToolList } from './tool-list-format';
+
+describe('formatToolList', () => {
+ test('joins with a locale conjunction so a checkbox label reads as prose', () => {
+ expect(formatToolList(['Claude', 'Cursor', 'Codex'], 'en')).toBe('Claude, Cursor, and Codex');
+ });
+
+ test('two tools take the conjunction with no serial comma', () => {
+ expect(formatToolList(['Claude', 'Cursor'], 'en')).toBe('Claude and Cursor');
+ });
+
+ test('a single tool renders bare', () => {
+ expect(formatToolList(['Claude'], 'en')).toBe('Claude');
+ });
+
+ test('an empty list renders empty rather than throwing', () => {
+ expect(formatToolList([], 'en')).toBe('');
+ });
+
+ test('an unset locale falls back to the runtime default instead of throwing', () => {
+ // `i18n.locale` is '' before Lingui activates, and Intl rejects an empty
+ // string rather than treating it as "unspecified".
+ expect(formatToolList(['Claude', 'Cursor'], '')).toContain('Claude');
+ });
+
+ test('the conjunction is localized, not hardcoded English', () => {
+ expect(formatToolList(['Claude', 'Cursor'], 'es')).toBe('Claude y Cursor');
+ });
+});
diff --git a/packages/app/src/lib/tool-list-format.ts b/packages/app/src/lib/tool-list-format.ts
new file mode 100644
index 000000000..afd3b6dcf
--- /dev/null
+++ b/packages/app/src/lib/tool-list-format.ts
@@ -0,0 +1,26 @@
+/**
+ * Shared formatting for the AI-tool lists both setup surfaces name in their
+ * checkbox labels (first-launch consent, create-project). Its own module rather
+ * than an export off either dialog: `McpConsentDialogBody` is deliberately
+ * `React.lazy()`-loaded, so importing a helper from it would drag the whole
+ * dialog into the eager bundle.
+ */
+
+/**
+ * Render tool labels as a locale-correct conjunction ("Claude, Cursor and
+ * Codex"). `Intl.ListFormat` rather than a joined string because the separator
+ * and the final conjunction differ per locale, and this list is the user's
+ * collapsed view of exactly which tools get written to.
+ *
+ * `locale` is required, matching the sibling `formatInstalls` — every call site
+ * passes `i18n.locale`, and one optional / one required signature for the same
+ * parameter invites a caller to wonder which is authoritative. It is normalized
+ * away when empty: `i18n.locale` is `''` before Lingui activates, and `Intl`
+ * rejects an empty string rather than treating it as unspecified.
+ */
+export function formatToolList(labels: readonly string[], locale: string): string {
+ return new Intl.ListFormat(locale || undefined, {
+ style: 'long',
+ type: 'conjunction',
+ }).format(labels);
+}
diff --git a/packages/app/src/locales/ar/messages.json b/packages/app/src/locales/ar/messages.json
index 763171621..c11773da5 100644
--- a/packages/app/src/locales/ar/messages.json
+++ b/packages/app/src/locales/ar/messages.json
@@ -355,20 +355,6 @@
"3pU8e-": ["تثبيت التبويب"],
"3qDsdf": ["مهارة التثبيت"],
"3ri_4O": ["يحافظ على توافق قاعدة معارفك مع Open Knowledge Format."],
- "3s-k7L": [
- [
- "0",
- "plural",
- {
- "zero": ["عرض ", "#", " أداة إضافية"],
- "one": ["عرض أداة إضافية واحدة"],
- "two": ["عرض أداتين إضافيتين"],
- "few": ["عرض ", "#", " أدوات إضافية"],
- "many": ["عرض ", "#", " أداةً إضافية"],
- "other": ["عرض ", "#", " أداة إضافية"]
- }
- ]
- ],
"3uSzeS": ["صور من أي مكان · يجب أن تكون الملفات غير المصورة داخل مساحة العمل"],
"3v62tp": ["شائع على skills.sh"],
"3wiBAs": ["أين ينبغي أن يعيش؟"],
@@ -742,7 +728,6 @@
["0"],
" — ذلك المجلد ليس هذه المهارة. أعد تسمية إحداهما أو احذفها للحل."
],
- "8IP5jY": ["تحصل كل أداة محددة على مدخل MCP لـ OpenKnowledge."],
"8Jm2EP": ["بدّل مستويات العناوين من 1 إلى 6."],
"8L0EPw": ["أو الصق رمزًا"],
"8OGgNM": ["افتح بحث المحرر المرئي أو أغلقه."],
@@ -1022,6 +1007,9 @@
"C0cll5": ["مثبَّت في ", ["0"], " — انقر للإزالة"],
"C0hoX6": ["مُدرج دائمًا"],
"C3Zc44": ["استكشاف المهارات"],
+ "C4HK3A": [
+ "يضيف إدخال MCP الخاص بـ OpenKnowledge ومهارة المشروع، حتى تتمكن كل أداة من قراءة هذه المستندات وتحديثها."
+ ],
"C62P6b": ["تحديد المحرر"],
"C7PKTv": ["إرسال التعليقات المحددة إلى المحادثة"],
"C7waqI": [
@@ -1743,6 +1731,7 @@
"L83DKQ": ["فشل التثبيت في PATH: ", ["summary"]],
"L8fKuc": ["شغّل هذا في طرفيتك"],
"L9Y9Om": ["أكثر الصفحات ارتباطًا"],
+ "LAR9ub": ["سيستبدل إدخال OpenKnowledge الحالي في ", ["0"]],
"LBv_TR": ["تحويل المواقع إلى روابط رمزية"],
"LBwUR6": ["انتهت مهلة تبديل الفرع — حاول فتح ", ["targetNoun"], " يدويًا."],
"LCUy_N": ["انتهت الجلسة"],
@@ -2025,6 +2014,9 @@
"OETysD": [
"لا توجد صفحات محورية بعد. تظهر المحاور عندما تتراكم للصفحات روابط واردة في الرسم البياني."
],
+ "OFIL46": [
+ "امنح أدوات الذكاء الاصطناعي التي تستخدمها إمكانية قراءة مشاريعك وتحديثها. يمكنك تغيير أي من هذا لاحقًا في الإعدادات > أدوات الذكاء الاصطناعي وواجهة سطر الأوامر."
+ ],
"OJ4VDl": [
"يعمل هذا الوكيل عبر ",
["0"],
@@ -2391,7 +2383,6 @@
"T3j-7o": ["مثال: تبديل المشاريع أثناء تشغيل مزامنة"],
"TBfy1C": ["تعذّر تشغيل Slidev."],
"TDeOpa": ["تراجع عن تعديلاتك أنت أو أعدها، مع إبقاء تعديلات المتعاونين والوكلاء كما هي."],
- "TEEPPF": ["سيستبدل مدخل OpenKnowledge الحالي"],
"TEJiKw": ["تحقق OpenKnowledge من ", ["subject"]],
"TKQ7K-": ["تثبيت"],
"TLIO3j": ["بدّل التنسيق المائل."],
@@ -3106,6 +3097,9 @@
". ثبّت أيًّا منها من الإضافة."
],
"aRI3Om": ["إغلاق النافذة"],
+ "aSVmqV": [
+ "لم يتم اكتشاف أي أداة ذكاء اصطناعي بعد. بعد تثبيت واحدة، اربطها من الإعدادات > هذا المشروع."
+ ],
"aWlJai": ["يجري بدء تدفق تسجيل الدخول"],
"aXkLuE": ["استجابة التطبيق مشوّهة"],
"a_2_uB": ["تعذّر إنشاء ", ["trimmed"], ": ", ["0"]],
@@ -3342,7 +3336,6 @@
"dVaXnv": ["لم تُحمَّل قواعد المحتوى بعد — أعد المحاولة بعد قليل"],
"dXCi3J": ["حدث عطل"],
"dXOolp": ["تجري مزامنة تغييراتك"],
- "dXWMFn": ["يزيل هذه المهارة من محرراتك."],
"darW93": [
[
"pageCount",
@@ -3417,6 +3410,7 @@
"ePK91l": ["تحرير"],
"eQkgKV": ["مثبَّت"],
"eQn7nI": ["جارٍ تحميل مشاركة الإعدادات"],
+ "eQsWFc": ["مهارة الوكيل"],
"eQxm13": ["فشل تحديث موقع المرفقات — ", ["detail"]],
"eRSPr9": ["فتح الطرفية"],
"eTL5y9": ["إظهار معاينة التحديد"],
@@ -3516,6 +3510,7 @@
"fhaxJL": ["فشل الاستنساخ — نص الاستجابة فارغ"],
"fiMNLl": ["لم تُحمَّل إعدادات المزامنة بعد — أعد المحاولة بعد قليل"],
"fj5YiP": ["ينزّل نسخة جديدة من GitHub."],
+ "fmwwju": ["يضيف اتصال MCP الخاص بـ OpenKnowledge إلى كل أداة."],
"fnlBN4": ["لم تُحمَّل إعدادات الموافقة التلقائية بعد — أعد المحاولة بعد قليل"],
"fo0VXg": ["إلغاء التثبيت"],
"fo6CCk": ["الجزء ", ["0"]],
@@ -3689,7 +3684,6 @@
"hmYd9H": [
"تتبّع الأشخاص والشركات التي تعمل معها والاجتماعات التي تعقدها معهم. مناسب لتذكّر من يكون أحدهم، وكيف تعرفه، وآخر ما تحدثتما عنه."
],
- "hnlGzG": ["التخطي الآن"],
"hnq8fP": ["إزالة ", ["label"], " من السياق"],
"hp8OtS": ["أُضيف"],
"hpDtUm": ["رابط ويكي خارجي"],
@@ -3809,6 +3803,7 @@
"j2kt1j": [
"افتح جلسة جديدة مع الذكاء الاصطناعي المفضّل لديك. مع تحديد نص، جهّزه هناك بدلًا من تشغيله."
],
+ "j4SmVY": ["ربط ", ["0"], " بهذا المشروع"],
"j5CWO4": ["محدَّث"],
"j9Vqvc": ["موروث"],
"j9tqwO": ["ليس في الشريط الجانبي"],
@@ -3833,6 +3828,7 @@
"jSO1X5": ["لا يمكن إعادة تسمية القوالب الموروثة هنا — أعد تسميتها في المجلد الذي يملكها."],
"jSY41h": ["استغرق «", ["docName"], "» وقتًا طويلًا. تحقق من اتصالك."],
"jTLWxh": ["ضمّن صفحة خارجية في إطار مضمَّن (مستندات، عروض، Figma، CodeSandbox)."],
+ "jWzDfC": ["جارٍ التحقق من أدوات الذكاء الاصطناعي المتوفّرة لديك"],
"jZ_E5M": ["فشل الجلب — راجع سجلات الخادم للتفاصيل."],
"j_Z6g7": ["الافتراضي المشترك"],
"jbSG8V": ["إظهار الملفات (", ["sidebarShortcutLabel"], ")"],
@@ -3977,9 +3973,6 @@
"kpqWwG": ["أُزيل هذا ", ["targetNoun"], " من الفرع <0>", ["branch"], "</0>."],
"ktspdq": ["لوّن وأشِر إلى الملفات التي بها مشكلات فحص أو روابط، دون فتحها."],
"ku0yLQ": ["تُضمَّن الصفحات أول مرة يحتاجها فيها بحث."],
- "kwLcuN": [
- "امنح أدوات الذكاء الاصطناعي التي تستخدمها إذن قراءة مشاريعك وتحديثها. اختر ما تريد إعداده أدناه، وغيّره في أي وقت من الإعدادات > أدوات الذكاء الاصطناعي وواجهة سطر الأوامر."
- ],
"kwNS82": ["لا يمكن استخدام هذه السمة حتى يُصلح الملف."],
"ky9Tuc": ["تحرير مصدر ", ["0"]],
"kzpkcn": ["تعذّر حفظ ذلك التعديل."],
@@ -4699,6 +4692,9 @@
"tHKQ6j": ["تجري إعادة التسمية"],
"tKlWWY": ["رموز تعبيرية"],
"tLX6LY": ["كتب OpenKnowledge مستندًا"],
+ "tLd2qa": [
+ "لم يتم اكتشاف أي أداة ذكاء اصطناعي بعد. بعد تثبيت واحدة، اربطها من الإعدادات > أدوات الذكاء الاصطناعي وواجهة سطر الأوامر."
+ ],
"tM5uzl": [
"عائلات المصدر والثقة ودورة الحياة: sources، generated، verified، status، وstale_after. كل حقل اختياري، والغياب ليس خطأ أبدًا."
],
@@ -5035,6 +5031,7 @@
"xEb24l": ["دعنا نتواصل معك بالبريد الإلكتروني"],
"xFQRu4": ["لديّ نسخة محلية بالفعل"],
"xG4tb6": ["فُقد الاتصال بـ «", ["docName"], "»."],
+ "xGVfLh": ["متابعة"],
"xHGM4U": ["إخفاء معاينة HTML"],
"xK291b": ["صفحتان (فردي)"],
"xKDsGS": ["الانتقال إلى السطر ", ["0"], " في ", ["1"]],
@@ -5141,6 +5138,9 @@
"ycIX5i": ["أنت خارج تسجيل الدخول — سجّل الدخول لاستئناف المزامنة."],
"yckias": ["إعادة تسمية الملف"],
"ydzS9x": ["خروج"],
+ "yeMZk8": [
+ "يضيف اتصال MCP الخاص بـ OpenKnowledge إلى كل أداة، بالإضافة إلى مهارة open-knowledge-discovery حتى يتعرّف وكلاؤك على مشاريع OpenKnowledge."
+ ],
"yf5auF": ["ابحث في القواعد بالمعرّف أو الاسم المستعار أو الاسم"],
"yf5r22": [
"نوع ",
diff --git a/packages/app/src/locales/ar/messages.po b/packages/app/src/locales/ar/messages.po
index d0bc36dcb..e23176312 100644
--- a/packages/app/src/locales/ar/messages.po
+++ b/packages/app/src/locales/ar/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, zero {لا مشكلات في # مستند.} one {لا مشك
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, zero {إضافات} one {إضافة} two {إضافتان} few {إضافات} many {إضافة} other {إضافة}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, zero {عرض # أداة إضافية} one {عرض أداة إضافية واحدة} two {عرض أداتين إضافيتين} few {عرض # أدوات إضافية} many {عرض # أداةً إضافية} other {عرض # أداة إضافية}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "يضيف كتلة مُدارة إلى"
msgid "Adds a managed block to {0}"
msgstr "يضيف كتلة مُدارة إلى {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "يضيف إدخال MCP الخاص بـ OpenKnowledge ومهارة المشروع، حتى تتمكن كل أداة من قراءة هذه المستندات وتحديثها."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "يضيف اتصال MCP الخاص بـ OpenKnowledge إلى كل أداة، بالإضافة إلى مهارة open-knowledge-discovery حتى يتعرّف وكلاؤك على مشاريع OpenKnowledge."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "يضيف اتصال MCP الخاص بـ OpenKnowledge إلى كل أداة."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "يضيف بيانات القياس وحالة الخادم ومعلومات وقت التشغيل عند توفرها. تُزال بيانات الاعتماد دائمًا؛ وتظهر أسماء المستندات، إن أُدرجت، كنص صريح (غير مخفي)."
@@ -1469,9 +1479,12 @@ msgstr "أُغلقت لوحة الوكيل لإبقاء Terminal سهل القر
msgid "Agent settings"
msgstr "إعدادات الوكيل"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "مهارة الوكيل"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "مهارات الوكيل"
@@ -1488,6 +1501,8 @@ msgstr "الوكلاء"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "إدارة أدوات الذكاء الاصطناعي متاحة فقط في تطبيق OpenKnowledge لسطح المكتب."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "مهارة التحقق"
msgid "Checking source"
msgstr "يجري التحقق من المصدر"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "جارٍ التحقق من أدوات الذكاء الاصطناعي المتوفّرة لديك"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "يجري التحقق من اتصالك بـ GitHub"
@@ -2925,12 +2944,15 @@ msgstr "الملفات المتعارضة: {0}."
msgid "Conflicts"
msgstr "التعارضات"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "اتصال"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "ربط {0} بهذا المشروع"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "اربط حساب GitHub لاستعراض مستودعاتك ومزامنتها."
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "اربط هذا المشروع بـ GitHub للمشاركة."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "الاتصال بأدوات الذكاء الاصطناعي"
@@ -3063,7 +3084,6 @@ msgstr "قواعد المحتوى"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "لم تُحمَّل قواعد المحتوى بعد — أعد المحاولة بعد قليل"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "تكلفة السياق"
@@ -3076,6 +3096,11 @@ msgstr "نافذة السياق:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "نافذة السياق: {percent}% مستخدَم، {usedLabel} من {sizeLabel} رمز"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "متابعة"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "المتابعة في المتصفح"
@@ -4661,10 +4686,6 @@ msgstr "مثال: تجمّد المحرر بعد لصق جدول كبير"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "تبني كل حزمة هيكل مشروعك بمجلدات وقوالب جاهزة."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "تحصل كل أداة محددة على مدخل MCP لـ OpenKnowledge."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "أدخل عنوان URL صالحًا، على سبيل المثال http
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "أدخل الأصل العام الذي يخدمه النفق الخاص بك، على سبيل المثال https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "الحقول"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "امنح أدوات الذكاء الاصطناعي التي تستخدمها إذن قراءة مشاريعك وتحديثها. تحديد مربع يُعدّه فورًا؛ وإلغاء التحديد يزيله."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "امنح أدوات الذكاء الاصطناعي التي تستخدمها إذن قراءة مشاريعك وتحديثها. اختر ما تريد إعداده أدناه، وغيّره في أي وقت من الإعدادات > أدوات الذكاء الاصطناعي وواجهة سطر الأوامر."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "امنح أدوات الذكاء الاصطناعي التي تستخدمها إمكانية قراءة مشاريعك وتحديثها. يمكنك تغيير أي من هذا لاحقًا في الإعدادات > أدوات الذكاء الاصطناعي وواجهة سطر الأوامر."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "إخفاء المتقدم"
msgid "Hide Agents"
msgstr "إخفاء الوكلاء"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "إخفاء التفاصيل"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "كيفية الدفع يدويًا"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "كيفية الإعداد"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "كيفية إعداد {0} (يفتح في المتصفح)"
@@ -6795,7 +6815,6 @@ msgstr "يجري التثبيت"
msgid "Installing {agentName}…"
msgstr "يجري تثبيت {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "يُثبَّت في"
@@ -7438,7 +7457,7 @@ msgstr "فشل الإصلاح التلقائي لـ MCP"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "لم يُحدَّد أي وكيل — المهارات محفوظة لك
msgid "No AI agents installed"
msgstr "لا توجد وكلاء ذكاء اصطناعي مثبَّتون"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "لم يتم اكتشاف أي أداة ذكاء اصطناعي بعد. بعد تثبيت واحدة، اربطها من الإعدادات > أدوات الذكاء الاصطناعي وواجهة سطر الأوامر."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "لم يتم اكتشاف أي أداة ذكاء اصطناعي بعد. بعد تثبيت واحدة، اربطها من الإعدادات > هذا المشروع."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "لم يتم العثور على أدوات الذكاء الاصطناعي. ثبّت واحدة لاستخدام هذه المهارة."
@@ -10179,11 +10207,6 @@ msgstr "أُزيل «{projectName}» من المشاريع الحديثة. لم
msgid "Removes from"
msgstr "يُزال من"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "يزيل هذه المهارة من محرراتك."
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "يجري الإزالة"
@@ -10338,6 +10361,11 @@ msgstr "هل تستبدل هذه النسخ بروابط رمزية إلى ال
msgid "Replace these symlinks with independent copies?"
msgstr "هل تستبدل هذه الروابط الرمزية بنسخ مستقلة؟"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "سيستبدل إدخال OpenKnowledge الحالي في {0}"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "أعدّه في صفحته الخاصة تحت «الإضافات»."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "اضبط المزامنة الافتراضية للمستخدمين الذين يفتحون هذا المشروع للمرة الأولى. يُودَع هذا الإعداد في مستودعك."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "إعداد {0}"
@@ -11405,9 +11433,7 @@ msgstr "إظهار لوحة المستند"
msgid "Show external URL nodes"
msgstr "إظهار عقد الروابط الخارجية"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "إظهار أقل"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "أصبحت المهارات مشتركة. أودع .ok/skills لمشاركتها مع فريقك."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "تُثبَّت المهارات في كل أداة ذكاء اصطناعي مكتشفة على هذا الجهاز، بغض النظر عن اتصالات MCP التي اخترتها أعلاه."
@@ -11655,11 +11679,6 @@ msgstr "skills.sh غير متاح؛ تُعرض نتائج GitHub (بلا ترت
msgid "Skip"
msgstr "تخطٍّ"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "التخطي الآن"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "ما الذي ينبغي أن يفعله الذكاء الاصطناعي
msgid "What the agent may read, write, and surface."
msgstr "ما يجوز للوكيل قراءته وكتابته وإظهاره."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "ما الذي يغيّره هذا"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "سيُنشأ في: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "سيُحفظ باسم <0>{sanitized}</0>."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "سيستبدل مدخل OpenKnowledge الحالي"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/bn/messages.json b/packages/app/src/locales/bn/messages.json
index 5f23e40d4..c70c60198 100644
--- a/packages/app/src/locales/bn/messages.json
+++ b/packages/app/src/locales/bn/messages.json
@@ -303,9 +303,6 @@
"3pU8e-": ["ট্যাব পিন করুন"],
"3qDsdf": ["দক্ষতা ইনস্টল করুন"],
"3ri_4O": ["আপনার জ্ঞানভান্ডারকে Open Knowledge Format-এর সঙ্গে সামঞ্জস্যপূর্ণ রাখে।"],
- "3s-k7L": [
- ["0", "plural", { "one": ["আরও ", "#", " টুল দেখান"], "other": ["আরও ", "#", " টুল দেখান"] }]
- ],
"3uSzeS": ["যেকোনো জায়গা থেকে ছবি · অ-ছবি ফাইলগুলি ওয়ার্কস্পেসে থাকতে হবে"],
"3v62tp": ["skills.sh-এ জনপ্রিয়"],
"3wiBAs": ["এটি কোথায় থাকবে?"],
@@ -607,7 +604,6 @@
["0"],
"-এ আছে — ওই ফোল্ডারটি এই স্কিল নয়। মীমাংসার জন্য দুটির একটির নাম বদলান বা মুছে দিন।"
],
- "8IP5jY": ["নির্বাচিত প্রতিটি টুল একটি OpenKnowledge MCP এন্ট্রি পায়।"],
"8Jm2EP": ["শিরোনামের স্তর ১ থেকে ৬ চালু বা বন্ধ করুন।"],
"8L0EPw": ["নয়তো একটি টোকেন পেস্ট করুন"],
"8OGgNM": ["দৃশ্য-সম্পাদকের খোঁজা খুলুন বা বন্ধ করুন।"],
@@ -851,6 +847,9 @@
"C0cll5": [["0"], "-এ ইনস্টল করা — সরাতে ক্লিক করুন"],
"C0hoX6": ["সবসময় অন্তর্ভুক্ত"],
"C3Zc44": ["স্কিল ঘুরে দেখুন"],
+ "C4HK3A": [
+ "OpenKnowledge-এর MCP এন্ট্রি এবং প্রকল্প Skill যোগ করে, যাতে প্রতিটি টুল এই নথিগুলি পড়তে ও আপডেট করতে পারে।"
+ ],
"C62P6b": ["সম্পাদকের নির্বাচন"],
"C7PKTv": ["চেক করা মন্তব্য আলাপে পাঠান"],
"C7waqI": [["0", "plural", { "one": ["#", " নথি"], "other": ["#", " নথি"] }]],
@@ -1474,6 +1473,7 @@
"L83DKQ": ["PATH ইনস্টল ব্যর্থ হয়েছে: ", ["summary"]],
"L8fKuc": ["এটি আপনার টার্মিনালে চালান"],
"L9Y9Om": ["সবচেয়ে বেশি লিংক করা পাতা"],
+ "LAR9ub": [["0"], "-এ থাকা বিদ্যমান OpenKnowledge এন্ট্রি প্রতিস্থাপন করবে"],
"LBv_TR": ["জায়গাগুলোকে সিমলিংকে রূপান্তর করুন"],
"LBwUR6": ["ব্রাঞ্চ বদলের সময় শেষ — ", ["targetNoun"], " হাতে খুলে দেখুন।"],
"LCUy_N": ["সেশন শেষ হয়েছে"],
@@ -1718,6 +1718,9 @@
"ODGGlj": ["ফোল্ডারের নাম: ", ["detail"]],
"OEEQKT": ["ভালো"],
"OETysD": ["এখনও কোনো হাব পাতা নেই। পাতাগুলোতে আগত গ্রাফ লিংক জমা হলে হাব দেখা দেয়।"],
+ "OFIL46": [
+ "আপনি যে AI টুল ব্যবহার করেন সেগুলিকে আপনার প্রকল্প পড়ার ও আপডেট করার অনুমতি দিন। এর সবকিছুই পরে Settings > AI tools & CLI-তে বদলানো যাবে।"
+ ],
"OJ4VDl": [
"এই এজেন্টটি ",
["0"],
@@ -2070,7 +2073,6 @@
"TDeOpa": [
"আপনার নিজের সম্পাদনা পূর্বাবস্থায় ফেরান বা পুনরায় করুন, সহযোগী ও এজেন্টদের সম্পাদনা যেমন আছে তেমন রেখে।"
],
- "TEEPPF": ["বিদ্যমান OpenKnowledge এন্ট্রির জায়গা নেবে"],
"TEJiKw": ["OpenKnowledge ", ["subject"], " পরীক্ষা করেছে"],
"TKQ7K-": ["ইনস্টল করুন"],
"TLIO3j": ["তির্যক বিন্যাস চালু বা বন্ধ করুন।"],
@@ -2663,6 +2665,9 @@
"। এগুলোর যেকোনোটি প্লাগইন থেকে ইনস্টল করুন।"
],
"aRI3Om": ["জানালা বন্ধ করুন"],
+ "aSVmqV": [
+ "এখনও কোনো AI টুল পাওয়া যায়নি। একটি ইনস্টল করার পর Settings > This project থেকে সেটি যুক্ত করুন।"
+ ],
"aWlJai": ["সাইন-ইন প্রক্রিয়া শুরু করা হচ্ছে"],
"aXkLuE": ["অ্যাপ্লাই প্রতিক্রিয়ার গঠন ত্রুটিপূর্ণ"],
"a_2_uB": [["trimmed"], " তৈরি করা যায়নি: ", ["0"]],
@@ -2856,7 +2861,6 @@
"dVaXnv": ["বিষয়বস্তুর নিয়ম এখনও লোড হয়নি — একটু পরে আবার চেষ্টা করুন"],
"dXCi3J": ["কিছু ভেঙে গেছে"],
"dXOolp": ["আপনার পরিবর্তন সিঙ্ক হচ্ছে"],
- "dXWMFn": ["এই স্কিলটি আপনার সম্পাদকগুলো থেকে সরিয়ে দেয়।"],
"darW93": [
[
"pageCount",
@@ -2909,6 +2913,7 @@
"ePK91l": ["সম্পাদনা করুন"],
"eQkgKV": ["ইনস্টল করা আছে"],
"eQn7nI": ["কনফিগ শেয়ারিং লোড হচ্ছে"],
+ "eQsWFc": ["Agent Skill"],
"eQxm13": ["সংযুক্তির জায়গা হালনাগাদ করতে ব্যর্থ — ", ["detail"]],
"eRSPr9": ["টার্মিনাল খুলুন"],
"eTL5y9": ["নির্বাচনের প্রাকদর্শন দেখান"],
@@ -2986,6 +2991,7 @@
"fhaxJL": ["ক্লোন ব্যর্থ — রেসপন্স বডি খালি"],
"fiMNLl": ["সিঙ্কের সেটিং এখনও লোড হয়নি — একটু পরে আবার চেষ্টা করুন"],
"fj5YiP": ["GitHub থেকে একটি নতুন অনুলিপি ডাউনলোড করে।"],
+ "fmwwju": ["প্রতিটিতে OpenKnowledge-এর MCP সংযোগ যোগ করে।"],
"fnlBN4": ["স্বয়ংক্রিয় অনুমোদনের সেটিং এখনও লোড হয়নি — একটু পরে আবার চেষ্টা করুন"],
"fo0VXg": ["আনইনস্টল করুন"],
"fo6CCk": ["প্যান ", ["0"]],
@@ -3139,7 +3145,6 @@
"hmYd9H": [
"আপনি যাদের ও যে কোম্পানিগুলোর সঙ্গে কাজ করেন এবং তাদের সঙ্গে আপনার বৈঠকগুলোর হিসাব রাখুন। কেউ কে, আপনি তাকে কীভাবে চেনেন, এবং শেষবার কী নিয়ে কথা হয়েছিল তা মনে রাখার জন্য ভালো।"
],
- "hnlGzG": ["আপাতত এড়িয়ে যান"],
"hnq8fP": [["label"], " প্রসঙ্গ থেকে সরান"],
"hp8OtS": ["যোগ করা হয়েছে"],
"hpDtUm": ["বাইরের উইকি লিংক"],
@@ -3256,6 +3261,7 @@
"j2kt1j": [
"আপনার পছন্দের AI দিয়ে একটি নতুন সেশন খুলুন। লেখা নির্বাচিত থাকলে সেটি চালানোর বদলে সেখানে প্রস্তুত রাখুন।"
],
+ "j4SmVY": [["0"], "-কে এই প্রকল্পের সঙ্গে যুক্ত করুন"],
"j5CWO4": ["হালনাগাদ"],
"j9Vqvc": ["উত্তরাধিকারসূত্রে পাওয়া"],
"j9tqwO": ["সাইডবারে নেই"],
@@ -3280,6 +3286,7 @@
"jSO1X5": ["উত্তরাধিকারসূত্রে পাওয়া টেমপ্লেটের নাম এখানে বদলানো যায় না — যে ফোল্ডারের এটি, সেখানেই নাম বদলান।"],
"jSY41h": ["\"", ["docName"], "\" খুব বেশি সময় নিয়েছে। আপনার সংযোগ দেখুন।"],
"jTLWxh": ["একটি বাইরের পাতা ইনলাইন iframe-এ এমবেড করুন (নথি, ডেমো, Figma, CodeSandbox)।"],
+ "jWzDfC": ["আপনার কোন AI টুল আছে তা যাচাই করা হচ্ছে"],
"jZ_E5M": ["আনা ব্যর্থ — বিস্তারিত জানতে সার্ভার লগ দেখুন।"],
"j_Z6g7": ["ভাগ করা ডিফল্ট"],
"jbSG8V": ["Files দেখান (", ["sidebarShortcutLabel"], ")"],
@@ -3408,9 +3415,6 @@
"kpqWwG": ["এই ", ["targetNoun"], " ব্রাঞ্চ <0>", ["branch"], "</0> থেকে সরানো হয়েছে।"],
"ktspdq": ["যেসব ফাইলে lint বা লিংকের সমস্যা আছে সেগুলো না খুলেই রঙ ও ব্যাজ দিন।"],
"ku0yLQ": ["কোনো অনুসন্ধানের প্রথমবার দরকার পড়লেই পাতাগুলো এমবেড করা হয়।"],
- "kwLcuN": [
- "আপনি যে AI টুলগুলো ব্যবহার করেন সেগুলোকে আপনার প্রকল্প পড়া ও হালনাগাদ করার অধিকার দিন। নিচে কী সেট আপ করবেন তা বাছুন, এবং Settings > AI tools & CLI-তে যেকোনো সময় বদলান।"
- ],
"kwNS82": ["ফাইলটি ঠিক না করা পর্যন্ত এই থিম ব্যবহার করা যাবে না।"],
"ky9Tuc": [["0"], " সোর্স সম্পাদনা করুন"],
"kzpkcn": ["ওই সম্পাদনাটি সংরক্ষণ করা যায়নি।"],
@@ -4006,6 +4010,9 @@
"tHKQ6j": ["নাম বদলানো হচ্ছে"],
"tKlWWY": ["ইমোজি"],
"tLX6LY": ["OpenKnowledge একটি নথি লিখেছে"],
+ "tLd2qa": [
+ "এখনও কোনো AI টুল পাওয়া যায়নি। একটি ইনস্টল করার পর Settings > AI tools & CLI থেকে সেটি যুক্ত করুন।"
+ ],
"tM5uzl": [
"মূল, বিশ্বাস, এবং জীবনচক্র পরিবারগুলি: sources, generated, verified, status, এবং stale_after। প্রতিটি ক্ষেত্র ঐচ্ছিক, এবং অনুপস্থিতি কখনও একটি ত্রুটি নয়।"
],
@@ -4280,6 +4287,7 @@
"xEb24l": ["আমাদের ইমেইলে যোগাযোগ করতে দিন"],
"xFQRu4": ["এটি আমার কাছে আগে থেকেই স্থানীয়ভাবে আছে"],
"xG4tb6": ["\"", ["docName"], "\"-এর সঙ্গে সংযোগ হারিয়েছে।"],
+ "xGVfLh": ["চালিয়ে যান"],
"xHGM4U": ["HTML প্রাকদর্শন লুকান"],
"xK291b": ["দুই পাতা (বিজোড়)"],
"xKDsGS": [["1"], "-এ লাইন ", ["0"], "-এ যান"],
@@ -4366,6 +4374,9 @@
"ycIX5i": ["আপনি সাইন আউট আছেন — সিঙ্ক আবার শুরু করতে সাইন ইন করুন।"],
"yckias": ["ফাইলের নাম বদলান"],
"ydzS9x": ["প্রস্থান করুন"],
+ "yeMZk8": [
+ "প্রতিটিতে OpenKnowledge-এর MCP সংযোগ যোগ করে, সঙ্গে open-knowledge-discovery Skill, যাতে আপনার এজেন্ট OpenKnowledge প্রকল্প চিনতে পারে।"
+ ],
"yf5auF": ["id, ডাকনাম বা নাম দিয়ে নিয়ম খুঁজুন"],
"yf5r22": [
["keyName"],
diff --git a/packages/app/src/locales/bn/messages.po b/packages/app/src/locales/bn/messages.po
index cebf38b7c..6b119d84e 100644
--- a/packages/app/src/locales/bn/messages.po
+++ b/packages/app/src/locales/bn/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {# নথিতে কোনো সমস্যা ন
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {প্লাগইন} other {প্লাগইন}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {আরও # টুল দেখান} other {আরও # টুল দেখান}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1410,6 +1405,21 @@ msgstr "এতে একটি পরিচালিত ব্লক যোগ
msgid "Adds a managed block to {0}"
msgstr "{0}-এ একটি পরিচালিত ব্লক যোগ করে"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "OpenKnowledge-এর MCP এন্ট্রি এবং প্রকল্প Skill যোগ করে, যাতে প্রতিটি টুল এই নথিগুলি পড়তে ও আপডেট করতে পারে।"
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "প্রতিটিতে OpenKnowledge-এর MCP সংযোগ যোগ করে, সঙ্গে open-knowledge-discovery Skill, যাতে আপনার এজেন্ট OpenKnowledge প্রকল্প চিনতে পারে।"
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "প্রতিটিতে OpenKnowledge-এর MCP সংযোগ যোগ করে।"
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "পাওয়া গেলে টেলিমেট্রি, সার্ভারের অবস্থা ও রানটাইম তথ্য যোগ করে। ক্রেডেনশিয়াল সবসময় সরিয়ে ফেলা হয়; নথির নাম, যদি অন্তর্ভুক্ত থাকে, সাদা লেখায় দেখা যায় (সরানো হয় না)।"
@@ -1465,9 +1475,12 @@ msgstr "Terminal পাঠযোগ্য রাখতে এজেন্ট প
msgid "Agent settings"
msgstr "এজেন্ট সেটিং"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Agent Skill"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Agent Skills"
@@ -1484,6 +1497,8 @@ msgstr "এজেন্ট"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI টুল ব্যবস্থাপনা শুধু OpenKnowledge ডেস্কটপ অ্যাপেই আছে।"
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2371,6 +2386,10 @@ msgstr "দক্ষতা পরীক্ষা করা"
msgid "Checking source"
msgstr "উৎস পরীক্ষা করা হচ্ছে"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "আপনার কোন AI টুল আছে তা যাচাই করা হচ্ছে"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "আপনার GitHub সংযোগ পরীক্ষা করা হচ্ছে"
@@ -2921,12 +2940,15 @@ msgstr "বিরোধপূর্ণ ফাইল: {0}।"
msgid "Conflicts"
msgstr "দ্বন্দ্ব"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "যুক্ত করুন"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "{0}-কে এই প্রকল্পের সঙ্গে যুক্ত করুন"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "আপনার রিপোজিটরি ব্রাউজ ও সিঙ্ক করতে একটি GitHub অ্যাকাউন্ট যুক্ত করুন।"
@@ -2960,7 +2982,6 @@ msgid "Connect this project to GitHub to share."
msgstr "ভাগ করতে এই প্রকল্পটি GitHub-এর সঙ্গে যুক্ত করুন।"
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "AI টুলের সঙ্গে যুক্ত করুন"
@@ -3059,7 +3080,6 @@ msgstr "বিষয়বস্তুর নিয়ম"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "বিষয়বস্তুর নিয়ম এখনও লোড হয়নি — একটু পরে আবার চেষ্টা করুন"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "কনটেক্সট খরচ"
@@ -3072,6 +3092,11 @@ msgstr "প্রসঙ্গ জানালা:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "প্রসঙ্গ জানালা: {percent}% ব্যবহৃত, {sizeLabel}-এর মধ্যে {usedLabel} টোকেন"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "চালিয়ে যান"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "ব্রাউজারে চালিয়ে যান"
@@ -4657,10 +4682,6 @@ msgstr "যেমন একটি বড় টেবিল পেস্ট ক
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "প্রতিটি প্যাক আপনার প্রকল্পকে তৈরি ফোল্ডার ও টেমপ্লেট দিয়ে কাঠামো দেয়।"
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "নির্বাচিত প্রতিটি টুল একটি OpenKnowledge MCP এন্ট্রি পায়।"
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5017,7 +5038,6 @@ msgstr "একটি বৈধ URL লিখুন, উদাহরণস্ব
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "আপনার টানেল যে পাবলিক অরিজিন পরিবেশন করে তা লিখুন, উদাহরণস্বরূপ https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5527,7 +5547,6 @@ msgstr "ফিল্ড"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5977,8 +5996,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "আপনি যে AI টুলগুলো ব্যবহার করেন সেগুলোকে আপনার প্রকল্প পড়া ও হালনাগাদ করার অধিকার দিন। একটি বাক্সে টিক দিলেই সেটি সঙ্গে সঙ্গে সেট হয়ে যায়; টিক সরালে সরে যায়।"
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "আপনি যে AI টুলগুলো ব্যবহার করেন সেগুলোকে আপনার প্রকল্প পড়া ও হালনাগাদ করার অধিকার দিন। নিচে কী সেট আপ করবেন তা বাছুন, এবং Settings > AI tools & CLI-তে যেকোনো সময় বদলান।"
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "আপনি যে AI টুল ব্যবহার করেন সেগুলিকে আপনার প্রকল্প পড়ার ও আপডেট করার অনুমতি দিন। এর সবকিছুই পরে Settings > AI tools & CLI-তে বদলানো যাবে।"
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6160,7 +6179,11 @@ msgstr "উন্নত লুকান"
msgid "Hide Agents"
msgstr "Agents লুকান"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "বিবরণ লুকান"
@@ -6302,14 +6325,11 @@ msgid "How to push manually"
msgstr "হাতে কীভাবে পাঠাবেন"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "কীভাবে সেট আপ করবেন"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "{0} কীভাবে সেট আপ করবেন (ব্রাউজারে খোলে)"
@@ -6791,7 +6811,6 @@ msgstr "ইনস্টল করা হচ্ছে"
msgid "Installing {agentName}…"
msgstr "{agentName} ইনস্টল করা হচ্ছে…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "যেখানে ইনস্টল হবে"
@@ -7434,7 +7453,7 @@ msgstr "MCP স্বয়ংক্রিয় মেরামত ব্যর
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8042,6 +8061,15 @@ msgstr "কোনো এজেন্ট নির্বাচিত হয়ন
msgid "No AI agents installed"
msgstr "কোনো AI এজেন্ট ইনস্টল করা নেই"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "এখনও কোনো AI টুল পাওয়া যায়নি। একটি ইনস্টল করার পর Settings > AI tools & CLI থেকে সেটি যুক্ত করুন।"
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "এখনও কোনো AI টুল পাওয়া যায়নি। একটি ইনস্টল করার পর Settings > This project থেকে সেটি যুক্ত করুন।"
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "কোনো AI টুল পাওয়া যায়নি। এই স্কিল ব্যবহার করতে একটি ইনস্টল করুন।"
@@ -10175,11 +10203,6 @@ msgstr "\"{projectName}\"-কে সাম্প্রতিক প্রকল
msgid "Removes from"
msgstr "যেখান থেকে সরানো হবে"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "এই স্কিলটি আপনার সম্পাদকগুলো থেকে সরিয়ে দেয়।"
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "সরানো হচ্ছে"
@@ -10334,6 +10357,11 @@ msgstr "এই অনুলিপিগুলো উৎসের সিমলি
msgid "Replace these symlinks with independent copies?"
msgstr "এই সিমলিংকগুলো স্বতন্ত্র অনুলিপি দিয়ে বদলাবেন?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "{0}-এ থাকা বিদ্যমান OpenKnowledge এন্ট্রি প্রতিস্থাপন করবে"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11209,7 +11237,7 @@ msgstr "Plugins-এর নিচে এর নিজের পাতায়
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "যাঁরা এই প্রকল্পটি প্রথমবার খুলবেন তাঁদের জন্য সিঙ্কের ডিফল্ট নির্ধারণ করুন। এই সেটিং আপনার রিপোজিটরিতে কমিট হয়।"
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "{0} সেট আপ করুন"
@@ -11401,9 +11429,7 @@ msgstr "নথি প্যানেল দেখান"
msgid "Show external URL nodes"
msgstr "বাইরের URL নোড দেখান"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "কম দেখান"
@@ -11622,8 +11648,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "স্কিল এখন ভাগ করা। আপনার দলের সঙ্গে ভাগ করতে .ok/skills কমিট করুন।"
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "স্কিল এই মেশিনে শনাক্ত হওয়া প্রতিটি AI টুলে ইনস্টল হয়, উপরে বেছে নেওয়া MCP সংযোগ যাই হোক না কেন।"
@@ -11651,11 +11675,6 @@ msgstr "skills.sh পাওয়া যাচ্ছে না; GitHub-এর
msgid "Skip"
msgstr "এড়িয়ে যান"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "আপাতত এড়িয়ে যান"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13989,11 +14008,13 @@ msgstr "AI-এর কী করা উচিত? (ঐচ্ছিক)"
msgid "What the agent may read, write, and surface."
msgstr "এজেন্ট কী পড়তে, লিখতে ও সামনে আনতে পারে।"
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "এটি কী বদলায়"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14087,11 +14108,6 @@ msgstr "এখানে তৈরি হবে: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "<0>{sanitized}</0> হিসেবে সংরক্ষিত হবে।"
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "বিদ্যমান OpenKnowledge এন্ট্রির জায়গা নেবে"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/en/messages.json b/packages/app/src/locales/en/messages.json
index d5491c595..be9561e6e 100644
--- a/packages/app/src/locales/en/messages.json
+++ b/packages/app/src/locales/en/messages.json
@@ -308,13 +308,6 @@
"3pU8e-": ["Pin tab"],
"3qDsdf": ["Install skill"],
"3ri_4O": ["Keeps your knowledge base aligned with the Open Knowledge Format."],
- "3s-k7L": [
- [
- "0",
- "plural",
- { "one": ["Show ", "#", " more tool"], "other": ["Show ", "#", " more tools"] }
- ]
- ],
"3uSzeS": ["Images from anywhere · non-image files must be in the workspace"],
"3v62tp": ["Popular on skills.sh"],
"3wiBAs": ["Where should it live?"],
@@ -626,7 +619,6 @@
["0"],
" — that folder is not this skill. Rename or delete one of the two to resolve."
],
- "8IP5jY": ["Each selected tool gets an OpenKnowledge MCP entry."],
"8Jm2EP": ["Toggle heading levels 1 through 6."],
"8L0EPw": ["or paste a token"],
"8OGgNM": ["Open or close visual-editor find."],
@@ -881,6 +873,9 @@
"C0cll5": ["Installed at ", ["0"], " — click to remove"],
"C0hoX6": ["Always included"],
"C3Zc44": ["Explore skills"],
+ "C4HK3A": [
+ "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+ ],
"C62P6b": ["Editor selection"],
"C7PKTv": ["Send checked comments to chat"],
"C7waqI": [["0", "plural", { "one": ["#", " document"], "other": ["#", " documents"] }]],
@@ -1515,6 +1510,7 @@
"L83DKQ": ["PATH install failed: ", ["summary"]],
"L8fKuc": ["Run this in your terminal"],
"L9Y9Om": ["Top linked pages"],
+ "LAR9ub": ["Replaces the existing OpenKnowledge entry in ", ["0"]],
"LBv_TR": ["Convert locations to symlinks"],
"LBwUR6": ["Branch switch timed out — try opening the ", ["targetNoun"], " manually."],
"LCUy_N": ["Session ended"],
@@ -1766,6 +1762,9 @@
"ODGGlj": ["Folder name: ", ["detail"]],
"OEEQKT": ["Good"],
"OETysD": ["No hub pages yet. Hubs appear once pages accumulate inbound graph links."],
+ "OFIL46": [
+ "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+ ],
"OJ4VDl": [
"This agent runs through ",
["0"],
@@ -2113,7 +2112,6 @@
"TDeOpa": [
"Undo or redo your own edits, leaving edits from collaborators and agents in place."
],
- "TEEPPF": ["Will replace existing OpenKnowledge entry"],
"TEJiKw": ["OpenKnowledge checked ", ["subject"]],
"TKQ7K-": ["Install"],
"TLIO3j": ["Toggle italic formatting."],
@@ -2716,6 +2714,9 @@
". Install any of them from the plugin."
],
"aRI3Om": ["Close Window"],
+ "aSVmqV": [
+ "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+ ],
"aWlJai": ["Starting sign-in flow"],
"aXkLuE": ["Malformed apply response"],
"a_2_uB": ["Couldn't create ", ["trimmed"], ": ", ["0"]],
@@ -2919,7 +2920,6 @@
"dVaXnv": ["Content rules not yet loaded — try again in a moment"],
"dXCi3J": ["Something broke"],
"dXOolp": ["Syncing your changes"],
- "dXWMFn": ["Removes this skill from your editors."],
"darW93": [
[
"pageCount",
@@ -2982,6 +2982,7 @@
"ePK91l": ["Edit"],
"eQkgKV": ["Installed"],
"eQn7nI": ["Loading config sharing"],
+ "eQsWFc": ["Agent Skill"],
"eQxm13": ["Failed to update attachment location — ", ["detail"]],
"eRSPr9": ["Open terminal"],
"eTL5y9": ["Show selection preview"],
@@ -3061,6 +3062,7 @@
"fhaxJL": ["Clone failed — empty response body"],
"fiMNLl": ["Sync settings not yet loaded — try again in a moment"],
"fj5YiP": ["Downloads a fresh copy from GitHub."],
+ "fmwwju": ["Adds OpenKnowledge's MCP connection to each."],
"fnlBN4": ["Auto-approve settings not yet loaded — try again in a moment"],
"fo0VXg": ["Uninstall"],
"fo6CCk": ["pane ", ["0"]],
@@ -3235,7 +3237,6 @@
"hmYd9H": [
"Keep track of the people and companies you work with and the meetings you have with them. Good for remembering who someone is, how you know them, and what you last talked about."
],
- "hnlGzG": ["Skip for now"],
"hnq8fP": ["Remove ", ["label"], " from context"],
"hp8OtS": ["Added"],
"hpDtUm": ["External wiki link"],
@@ -3352,6 +3353,7 @@
"j2kt1j": [
"Open a new session with your preferred AI. With text selected, stage it there instead of running it."
],
+ "j4SmVY": ["Connect ", ["0"], " to this project"],
"j5CWO4": ["Up to date"],
"j9Vqvc": ["inherited"],
"j9tqwO": ["Not in sidebar"],
@@ -3376,6 +3378,7 @@
"jSO1X5": ["Inherited templates can't be renamed here — rename it in the folder that owns it."],
"jSY41h": ["\"", ["docName"], "\" took too long. Check your connection."],
"jTLWxh": ["Embed an external page in an inline iframe (docs, demos, Figma, CodeSandbox)."],
+ "jWzDfC": ["Checking which AI tools you have"],
"jZ_E5M": ["Fetch failed — check the server logs for details."],
"j_Z6g7": ["Shared default"],
"jbSG8V": ["Show Files (", ["sidebarShortcutLabel"], ")"],
@@ -3507,9 +3510,6 @@
"kpqWwG": ["This ", ["targetNoun"], " was removed from branch <0>", ["branch"], "</0>."],
"ktspdq": ["Tint and badge files that have lint or link problems, without opening them."],
"ku0yLQ": ["Pages are embedded the first time a search needs them."],
- "kwLcuN": [
- "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
- ],
"kwNS82": ["This theme can’t be used until the file is fixed."],
"ky9Tuc": ["Edit ", ["0"], " source"],
"kzpkcn": ["Couldn't save that edit."],
@@ -4134,6 +4134,9 @@
"tHKQ6j": ["Renaming"],
"tKlWWY": ["Emoji"],
"tLX6LY": ["OpenKnowledge wrote a document"],
+ "tLd2qa": [
+ "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+ ],
"tM5uzl": [
"The provenance, trust, and lifecycle families: sources, generated, verified, status, and stale_after. Every field is optional, and absence is never an error."
],
@@ -4428,6 +4431,7 @@
"xEb24l": ["Let us follow up by email"],
"xFQRu4": ["I already have it locally"],
"xG4tb6": ["Lost connection to \"", ["docName"], "\"."],
+ "xGVfLh": ["Continue"],
"xHGM4U": ["Hide HTML preview"],
"xK291b": ["Two-page (odd)"],
"xKDsGS": ["Go to line ", ["0"], " in ", ["1"]],
@@ -4519,6 +4523,9 @@
"ycIX5i": ["You're signed out — sign in to resume syncing."],
"yckias": ["Rename file"],
"ydzS9x": ["Exit"],
+ "yeMZk8": [
+ "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+ ],
"yf5auF": ["Search rules by id, alias, or name"],
"yf5r22": [
["keyName"],
diff --git a/packages/app/src/locales/en/messages.po b/packages/app/src/locales/en/messages.po
index 33c0acac6..09f49d384 100644
--- a/packages/app/src/locales/en/messages.po
+++ b/packages/app/src/locales/en/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {No problems across # document.} other {No problems acro
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {Plugin} other {Plugins}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {Show # more tool} other {Show # more tools}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "Adds a managed block to"
msgid "Adds a managed block to {0}"
msgstr "Adds a managed block to {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "Adds OpenKnowledge's MCP connection to each."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
@@ -1469,9 +1479,12 @@ msgstr "Agent panel closed to keep Terminal readable."
msgid "Agent settings"
msgstr "Agent settings"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Agent Skill"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Agent Skills"
@@ -1488,6 +1501,8 @@ msgstr "Agents"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI tool management is only available in the OpenKnowledge desktop app."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "Checking skill"
msgid "Checking source"
msgstr "Checking source"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "Checking which AI tools you have"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "Checking your GitHub connection"
@@ -2925,12 +2944,15 @@ msgstr "Conflicting files: {0}."
msgid "Conflicts"
msgstr "Conflicts"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "Connect"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "Connect {0} to this project"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "Connect a GitHub account to browse and sync your repositories."
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "Connect this project to GitHub to share."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "Connect to AI tools"
@@ -3063,7 +3084,6 @@ msgstr "Content rules"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "Content rules not yet loaded — try again in a moment"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "Context cost"
@@ -3076,6 +3096,11 @@ msgstr "Context window:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "Continue"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "Continue in browser"
@@ -4661,10 +4686,6 @@ msgstr "e.g. The editor froze after I pasted a large table"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "Each pack scaffolds your project with ready-made folders and templates."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "Each selected tool gets an OpenKnowledge MCP entry."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "Enter a valid URL, for example https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "Enter the public origin your tunnel serves, for example https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "Fields"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "Give the AI tools you use access to read and update your projects. Checking a box sets it up right away; unchecking removes it."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "Hide advanced"
msgid "Hide Agents"
msgstr "Hide Agents"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "Hide details"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "How to push manually"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "How to set up"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "How to set up {0} (opens in browser)"
@@ -6795,7 +6815,6 @@ msgstr "Installing"
msgid "Installing {agentName}…"
msgstr "Installing {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "Installs to"
@@ -7438,7 +7457,7 @@ msgstr "MCP auto-repair failed"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "No agents selected — the skills are saved but not installed anywhere."
msgid "No AI agents installed"
msgstr "No AI agents installed"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "No AI tools detected. Install one to use this skill."
@@ -10179,11 +10207,6 @@ msgstr "Removed \"{projectName}\" from recent projects. Its folder no longer exi
msgid "Removes from"
msgstr "Removes from"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "Removes this skill from your editors."
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "Removing"
@@ -10338,6 +10361,11 @@ msgstr "Replace these copies with symlinks to the source?"
msgid "Replace these symlinks with independent copies?"
msgstr "Replace these symlinks with independent copies?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "Replaces the existing OpenKnowledge entry in {0}"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "Set it up on its own page under Plugins."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "Set up {0}"
@@ -11405,9 +11433,7 @@ msgstr "Show document panel"
msgid "Show external URL nodes"
msgstr "Show external URL nodes"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "Show fewer"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "Skills are now shared. Commit .ok/skills to share them with your team."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
@@ -11655,11 +11679,6 @@ msgstr "skills.sh is unavailable; showing GitHub results (no install ranking)."
msgid "Skip"
msgstr "Skip"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "Skip for now"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "What should the AI do? (optional)"
msgid "What the agent may read, write, and surface."
msgstr "What the agent may read, write, and surface."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "What this changes"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "Will be created at: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "Will be saved as <0>{sanitized}</0>."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "Will replace existing OpenKnowledge entry"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/es/messages.json b/packages/app/src/locales/es/messages.json
index a96e33260..ca1c29d92 100644
--- a/packages/app/src/locales/es/messages.json
+++ b/packages/app/src/locales/es/messages.json
@@ -357,17 +357,6 @@
"3pU8e-": ["Fijar pestaña"],
"3qDsdf": ["Instalar habilidad"],
"3ri_4O": ["Mantiene tu base de conocimientos alineada con el Open Knowledge Format."],
- "3s-k7L": [
- [
- "0",
- "plural",
- {
- "one": ["Mostrar ", "#", " herramienta más"],
- "many": ["Mostrar ", "#", " herramientas más"],
- "other": ["Mostrar ", "#", " herramientas más"]
- }
- ]
- ],
"3uSzeS": [
"Imágenes desde cualquier lugar · los archivos que no son imágenes deben estar en el espacio de trabajo"
],
@@ -744,7 +733,6 @@
["0"],
" vive una habilidad DISTINTA con un nombre parecido a esta; esa carpeta no es esta habilidad. Cambia el nombre o elimina una de las dos para resolverlo."
],
- "8IP5jY": ["Cada herramienta seleccionada recibe una entrada MCP de OpenKnowledge."],
"8Jm2EP": ["Activa o desactiva los niveles de encabezado del 1 al 6."],
"8L0EPw": ["o pega un token"],
"8OGgNM": ["Abre o cierra la búsqueda del editor visual."],
@@ -1032,6 +1020,9 @@
"C0cll5": ["Instalado en ", ["0"], ": haz clic para quitarlo"],
"C0hoX6": ["Siempre incluido"],
"C3Zc44": ["Explorar habilidades"],
+ "C4HK3A": [
+ "Añade una entrada MCP de OpenKnowledge y la habilidad del proyecto, para que cada herramienta pueda leer y actualizar estos documentos."
+ ],
"C62P6b": ["Selección del editor"],
"C7PKTv": ["Enviar los comentarios marcados al chat"],
"C7waqI": [
@@ -1770,6 +1761,7 @@
"L83DKQ": ["Falló la instalación en PATH: ", ["summary"]],
"L8fKuc": ["Ejecuta esto en tu terminal"],
"L9Y9Om": ["Páginas más enlazadas"],
+ "LAR9ub": ["Reemplaza la entrada de OpenKnowledge existente en ", ["0"]],
"LBv_TR": ["Convertir las ubicaciones en enlaces simbólicos"],
"LBwUR6": [
"Se agotó el tiempo al cambiar de rama: prueba a abrir ",
@@ -2050,6 +2042,9 @@
"OETysD": [
"Todavía no hay páginas índice. Los índices aparecen cuando las páginas acumulan enlaces entrantes en el grafo."
],
+ "OFIL46": [
+ "Da a las herramientas de IA que usas acceso para leer y actualizar tus proyectos. Puedes cambiar todo esto más tarde en Ajustes > Herramientas de IA y CLI."
+ ],
"OJ4VDl": [
"Este agente funciona con ",
["0"],
@@ -2426,7 +2421,6 @@
"TDeOpa": [
"Deshaz o rehaz tus propias ediciones, dejando intactas las de colaboradores y agentes."
],
- "TEEPPF": ["Sustituirá la entrada existente de OpenKnowledge"],
"TEJiKw": ["OpenKnowledge revisó ", ["subject"]],
"TKQ7K-": ["Instalar"],
"TLIO3j": ["Activa o desactiva la cursiva."],
@@ -3146,6 +3140,9 @@
". Puedes instalar cualquiera de ellas desde el complemento."
],
"aRI3Om": ["Cerrar ventana"],
+ "aSVmqV": [
+ "Todavía no se ha detectado ninguna herramienta de IA. Cuando instales una, conéctala desde Ajustes > Este proyecto."
+ ],
"aWlJai": ["Iniciando el flujo de inicio de sesión"],
"aXkLuE": ["Respuesta de aplicación con formato incorrecto"],
"a_2_uB": ["No se pudo crear ", ["trimmed"], ": ", ["0"]],
@@ -3392,7 +3389,6 @@
"dVaXnv": ["Las reglas de contenido aún no se han cargado; inténtalo de nuevo en un momento"],
"dXCi3J": ["Algo se rompió"],
"dXOolp": ["Sincronizando tus cambios"],
- "dXWMFn": ["Quita esta habilidad de tus editores."],
"darW93": [
[
"pageCount",
@@ -3470,6 +3466,7 @@
"ePK91l": ["Editar"],
"eQkgKV": ["Instalado"],
"eQn7nI": ["Cargando uso compartido de la configuración"],
+ "eQsWFc": ["Habilidad de agente"],
"eQxm13": ["No se pudo actualizar la ubicación de los adjuntos — ", ["detail"]],
"eRSPr9": ["Abrir la terminal"],
"eTL5y9": ["Mostrar la vista previa de la selección"],
@@ -3573,6 +3570,7 @@
"Los ajustes de sincronización aún no se han cargado; inténtalo de nuevo en un momento"
],
"fj5YiP": ["Descarga una copia nueva desde GitHub."],
+ "fmwwju": ["Añade la conexión MCP de OpenKnowledge a cada una."],
"fnlBN4": [
"Los ajustes de aprobación automática aún no se han cargado; inténtalo de nuevo en un momento"
],
@@ -3764,7 +3762,6 @@
"hmYd9H": [
"Lleva la cuenta de las personas y empresas con las que trabajas y de las reuniones que tienes con ellas. Va bien para recordar quién es alguien, de qué la conoces y de qué hablasteis la última vez."
],
- "hnlGzG": ["Omitir por ahora"],
"hnq8fP": ["Quitar ", ["label"], " del contexto"],
"hp8OtS": ["Añadido"],
"hpDtUm": ["Enlace wiki externo"],
@@ -3897,6 +3894,7 @@
"j2kt1j": [
"Abre una sesión nueva con tu IA preferida. Con texto seleccionado, lo prepara ahí en lugar de ejecutarlo."
],
+ "j4SmVY": ["Conectar ", ["0"], " a este proyecto"],
"j5CWO4": ["Al día"],
"j9Vqvc": ["heredado"],
"j9tqwO": ["No está en la barra lateral"],
@@ -3925,6 +3923,7 @@
"jTLWxh": [
"Incrusta una página externa en un iframe en línea (documentación, demos, Figma, CodeSandbox)."
],
+ "jWzDfC": ["Comprobando qué herramientas de IA tienes"],
"jZ_E5M": ["Falló el fetch: consulta los registros del servidor para ver los detalles."],
"j_Z6g7": ["Valor compartido por defecto"],
"jbSG8V": ["Mostrar Archivos (", ["sidebarShortcutLabel"], ")"],
@@ -4071,9 +4070,6 @@
"ku0yLQ": [
"Las páginas se convierten en embeddings la primera vez que una búsqueda las necesita."
],
- "kwLcuN": [
- "Da a las herramientas de IA que usas acceso para leer y actualizar tus proyectos. Elige abajo qué configurar; puedes cambiarlo cuando quieras en Ajustes > Herramientas de IA y CLI."
- ],
"kwNS82": ["Este tema no se puede usar hasta que se corrija el archivo."],
"ky9Tuc": ["Editar el código de ", ["0"]],
"kzpkcn": ["No se pudo guardar esa edición."],
@@ -4775,6 +4771,9 @@
"tHKQ6j": ["Renombrando"],
"tKlWWY": ["Emoji"],
"tLX6LY": ["OpenKnowledge escribió un documento"],
+ "tLd2qa": [
+ "Todavía no se ha detectado ninguna herramienta de IA. Cuando instales una, conéctala desde Ajustes > Herramientas de IA y CLI."
+ ],
"tM5uzl": [
"Las familias de procedencia, confianza y ciclo de vida: sources, generated, verified, status y stale_after. Cada campo es opcional y la ausencia nunca es un error."
],
@@ -5116,6 +5115,7 @@
"xEb24l": ["Permítenos escribirte por correo"],
"xFQRu4": ["Ya lo tengo en local"],
"xG4tb6": ["Se perdió la conexión con \"", ["docName"], "\"."],
+ "xGVfLh": ["Continuar"],
"xHGM4U": ["Ocultar la vista previa del HTML"],
"xK291b": ["Dos páginas (impares)"],
"xKDsGS": ["Ir a la línea ", ["0"], " de ", ["1"]],
@@ -5218,6 +5218,9 @@
"ycIX5i": ["Has cerrado sesión: inicia sesión para reanudar la sincronización."],
"yckias": ["Renombrar el archivo"],
"ydzS9x": ["Salir"],
+ "yeMZk8": [
+ "Añade la conexión MCP de OpenKnowledge a cada una, más la habilidad open-knowledge-discovery para que tus agentes reconozcan los proyectos de OpenKnowledge."
+ ],
"yf5auF": ["Busca reglas por id, alias o nombre"],
"yf5r22": [
"Tipo de ",
diff --git a/packages/app/src/locales/es/messages.po b/packages/app/src/locales/es/messages.po
index 40f7825fd..8e85581ee 100644
--- a/packages/app/src/locales/es/messages.po
+++ b/packages/app/src/locales/es/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {Sin problemas en # documento.} many {Sin problemas en #
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {Complemento} many {Complementos} other {Complementos}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {Mostrar # herramienta más} many {Mostrar # herramientas más} other {Mostrar # herramientas más}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "Añade un bloque gestionado a"
msgid "Adds a managed block to {0}"
msgstr "Añade un bloque gestionado a {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "Añade una entrada MCP de OpenKnowledge y la habilidad del proyecto, para que cada herramienta pueda leer y actualizar estos documentos."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "Añade la conexión MCP de OpenKnowledge a cada una, más la habilidad open-knowledge-discovery para que tus agentes reconozcan los proyectos de OpenKnowledge."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "Añade la conexión MCP de OpenKnowledge a cada una."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "Añade telemetría, estado del servidor e información de ejecución cuando está disponible. Las credenciales siempre se eliminan; los nombres de documento, si se incluyen, aparecen en texto plano (sin ocultar)."
@@ -1469,9 +1479,12 @@ msgstr "Se cerró el panel de agentes para mantener Terminal legible."
msgid "Agent settings"
msgstr "Ajustes del agente"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Habilidad de agente"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Habilidades de agente"
@@ -1488,6 +1501,8 @@ msgstr "Agentes"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "La gestión de herramientas de IA solo está disponible en la aplicación de escritorio de OpenKnowledge."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "Comprobando habilidad"
msgid "Checking source"
msgstr "Comprobando el origen"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "Comprobando qué herramientas de IA tienes"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "Comprobando tu conexión con GitHub"
@@ -2925,12 +2944,15 @@ msgstr "Archivos en conflicto: {0}."
msgid "Conflicts"
msgstr "Conflictos"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "Conectar"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "Conectar {0} a este proyecto"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "Conecta una cuenta de GitHub para explorar y sincronizar tus repositorios."
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "Conecta este proyecto a GitHub para compartirlo."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "Conectar con herramientas de IA"
@@ -3063,7 +3084,6 @@ msgstr "Reglas de contenido"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "Las reglas de contenido aún no se han cargado; inténtalo de nuevo en un momento"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "Coste de contexto"
@@ -3076,6 +3096,11 @@ msgstr "Ventana de contexto:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "Ventana de contexto: {percent}% usado, {usedLabel} de {sizeLabel} tokens"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "Continuar"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "Continuar en el navegador"
@@ -4661,10 +4686,6 @@ msgstr "p. ej. El editor se quedó bloqueado al pegar una tabla grande"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "Cada paquete prepara tu proyecto con carpetas y plantillas ya hechas."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "Cada herramienta seleccionada recibe una entrada MCP de OpenKnowledge."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "Introduce una URL válida, por ejemplo https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "Introduce el origen público que sirve tu túnel, por ejemplo https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "Campos"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "Da a las herramientas de IA que usas acceso para leer y actualizar tus proyectos. Al marcar una casilla se configura al momento; al desmarcarla se quita."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "Da a las herramientas de IA que usas acceso para leer y actualizar tus proyectos. Elige abajo qué configurar; puedes cambiarlo cuando quieras en Ajustes > Herramientas de IA y CLI."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "Da a las herramientas de IA que usas acceso para leer y actualizar tus proyectos. Puedes cambiar todo esto más tarde en Ajustes > Herramientas de IA y CLI."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "Ocultar opciones avanzadas"
msgid "Hide Agents"
msgstr "Ocultar Agentes"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "Ocultar detalles"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "Cómo hacer push manualmente"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "Cómo configurarlo"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "Cómo configurar {0} (se abre en el navegador)"
@@ -6795,7 +6815,6 @@ msgstr "Instalando"
msgid "Installing {agentName}…"
msgstr "Instalando {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "Se instala en"
@@ -7438,7 +7457,7 @@ msgstr "Falló la reparación automática de MCP"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "No hay agentes seleccionados: las habilidades se guardan pero no se inst
msgid "No AI agents installed"
msgstr "No hay agentes de IA instalados"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "Todavía no se ha detectado ninguna herramienta de IA. Cuando instales una, conéctala desde Ajustes > Herramientas de IA y CLI."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "Todavía no se ha detectado ninguna herramienta de IA. Cuando instales una, conéctala desde Ajustes > Este proyecto."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "No se detectaron herramientas de IA. Instala una para usar esta habilidad."
@@ -10179,11 +10207,6 @@ msgstr "Se quitó \"{projectName}\" de los proyectos recientes. Su carpeta ya no
msgid "Removes from"
msgstr "Se elimina de"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "Quita esta habilidad de tus editores."
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "Quitando"
@@ -10338,6 +10361,11 @@ msgstr "¿Reemplazar estas copias por enlaces simbólicos al origen?"
msgid "Replace these symlinks with independent copies?"
msgstr "¿Reemplazar estos enlaces simbólicos por copias independientes?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "Reemplaza la entrada de OpenKnowledge existente en {0}"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "Configúralo en su propia página, bajo Complementos."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "Define la sincronización por defecto para quienes abran este proyecto por primera vez. Este ajuste se guarda en tu repositorio."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "Configurar {0}"
@@ -11405,9 +11433,7 @@ msgstr "Mostrar el panel del documento"
msgid "Show external URL nodes"
msgstr "Mostrar los nodos de URL externas"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "Mostrar menos"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "Las habilidades ahora se comparten. Haz commit de .ok/skills para compartirlas con tu equipo."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "Las habilidades se instalan en todas las herramientas de IA detectadas en este equipo, con independencia de las conexiones MCP que elegiste arriba."
@@ -11655,11 +11679,6 @@ msgstr "skills.sh no está disponible; se muestran resultados de GitHub (sin cla
msgid "Skip"
msgstr "Omitir"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "Omitir por ahora"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "¿Qué debería hacer la IA? (opcional)"
msgid "What the agent may read, write, and surface."
msgstr "Lo que el agente puede leer, escribir y mostrar."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "Qué cambia esto"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "Se creará en: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "Se guardará como <0>{sanitized}</0>."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "Sustituirá la entrada existente de OpenKnowledge"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/fr/messages.json b/packages/app/src/locales/fr/messages.json
index e95be5fb9..2e20e077c 100644
--- a/packages/app/src/locales/fr/messages.json
+++ b/packages/app/src/locales/fr/messages.json
@@ -371,17 +371,6 @@
"3pU8e-": ["Épingler l'onglet"],
"3qDsdf": ["Installer la compétence"],
"3ri_4O": ["Maintient votre base de connaissances alignée sur l’Open Knowledge Format."],
- "3s-k7L": [
- [
- "0",
- "plural",
- {
- "one": ["Afficher ", "#", " outil de plus"],
- "many": ["Afficher ", "#", " outils de plus"],
- "other": ["Afficher ", "#", " outils de plus"]
- }
- ]
- ],
"3uSzeS": [
"Images de n'importe où · les fichiers non-image doivent être dans l'espace de travail"
],
@@ -768,7 +757,6 @@
["0"],
" — ce dossier n'est pas cette compétence. Renommez ou supprimez l'une des deux pour résoudre le problème."
],
- "8IP5jY": ["Chaque outil sélectionné reçoit une entrée MCP OpenKnowledge."],
"8Jm2EP": ["Bascule les niveaux de titre 1 à 6."],
"8L0EPw": ["ou collez un jeton"],
"8OGgNM": ["Ouvre ou ferme la recherche de l'éditeur visuel."],
@@ -1056,6 +1044,9 @@
"C0cll5": ["Installé dans ", ["0"], " — cliquez pour retirer"],
"C0hoX6": ["Toujours inclus"],
"C3Zc44": ["Explorer les compétences"],
+ "C4HK3A": [
+ "Ajoute une entrée MCP OpenKnowledge et la compétence du projet, pour que chaque outil puisse lire et mettre à jour ces documents."
+ ],
"C62P6b": ["Sélection de l'éditeur"],
"C7PKTv": ["Envoyer les commentaires cochés dans la discussion"],
"C7waqI": [
@@ -1797,6 +1788,7 @@
"L83DKQ": ["L'installation dans PATH a échoué : ", ["summary"]],
"L8fKuc": ["Exécutez ceci dans votre terminal"],
"L9Y9Om": ["Pages les plus liées"],
+ "LAR9ub": ["Remplace l'entrée OpenKnowledge existante dans ", ["0"]],
"LBv_TR": ["Convertir les emplacements en liens symboliques"],
"LBwUR6": [
"Le changement de branche a expiré — essayez d'ouvrir ",
@@ -2075,6 +2067,9 @@
"OETysD": [
"Pas encore de pages pivots. Les pivots apparaissent une fois que les pages accumulent des liens entrants dans le graphe."
],
+ "OFIL46": [
+ "Donnez aux outils IA que vous utilisez l'accès en lecture et en écriture à vos projets. Vous pourrez tout modifier plus tard dans Paramètres > Outils IA et CLI."
+ ],
"OJ4VDl": [
"Cet agent fonctionne via ",
["0"],
@@ -2453,7 +2448,6 @@
"TDeOpa": [
"Annule ou rétablit vos propres modifications, en laissant en place celles des collaborateurs et des agents."
],
- "TEEPPF": ["Remplacera l'entrée OpenKnowledge existante"],
"TEJiKw": ["OpenKnowledge a vérifié ", ["subject"]],
"TKQ7K-": ["Installer"],
"TLIO3j": ["Bascule la mise en forme italique."],
@@ -3176,6 +3170,9 @@
". Installez-en n'importe laquelle depuis le plugin."
],
"aRI3Om": ["Fermer la fenêtre"],
+ "aSVmqV": [
+ "Aucun outil IA détecté pour l'instant. Une fois que vous en installez un, connectez-le depuis Paramètres > Ce projet."
+ ],
"aWlJai": ["Démarrage du processus de connexion"],
"aXkLuE": ["Réponse d'application mal formée"],
"a_2_uB": ["Impossible de créer ", ["trimmed"], " : ", ["0"]],
@@ -3425,7 +3422,6 @@
"dVaXnv": ["Les règles de contenu ne sont pas encore chargées — réessayez dans un instant"],
"dXCi3J": ["Quelque chose s'est cassé"],
"dXOolp": ["Synchronisation de vos modifications"],
- "dXWMFn": ["Retire cette compétence de vos éditeurs."],
"darW93": [
[
"pageCount",
@@ -3501,6 +3497,7 @@
"ePK91l": ["Modifier"],
"eQkgKV": ["Installé"],
"eQn7nI": ["Chargement du partage de la configuration"],
+ "eQsWFc": ["Compétence d'agent"],
"eQxm13": ["Échec de la mise à jour de l'emplacement des pièces jointes — ", ["detail"]],
"eRSPr9": ["Ouvrir le terminal"],
"eTL5y9": ["Afficher l'aperçu de la sélection"],
@@ -3602,6 +3599,7 @@
"Les paramètres de synchronisation ne sont pas encore chargés — réessayez dans un instant"
],
"fj5YiP": ["Télécharge une copie neuve depuis GitHub."],
+ "fmwwju": ["Ajoute la connexion MCP d'OpenKnowledge à chacun."],
"fnlBN4": [
"Les paramètres d'approbation automatique ne sont pas encore chargés — réessayez dans un instant"
],
@@ -3791,7 +3789,6 @@
"hmYd9H": [
"Gardez la trace des personnes et des entreprises avec lesquelles vous travaillez et des réunions que vous avez avec elles. Utile pour se rappeler qui est quelqu'un, comment vous le connaissez, et de quoi vous avez parlé la dernière fois."
],
- "hnlGzG": ["Passer pour l'instant"],
"hnq8fP": ["Retirer ", ["label"], " du contexte"],
"hp8OtS": ["Ajouté"],
"hpDtUm": ["Lien wiki externe"],
@@ -3926,6 +3923,7 @@
"j2kt1j": [
"Ouvre une nouvelle session avec l'IA de votre choix. Si du texte est sélectionné, il y est préparé au lieu d'être exécuté."
],
+ "j4SmVY": ["Connecter ", ["0"], " à ce projet"],
"j5CWO4": ["À jour"],
"j9Vqvc": ["hérité"],
"j9tqwO": ["Pas dans la barre latérale"],
@@ -3954,6 +3952,7 @@
"jTLWxh": [
"Intègre une page externe dans une iframe en ligne (docs, démos, Figma, CodeSandbox)."
],
+ "jWzDfC": ["Vérification des outils IA dont vous disposez"],
"jZ_E5M": ["Échec de la récupération — consultez les journaux du serveur pour les détails."],
"j_Z6g7": ["Valeur par défaut partagée"],
"jbSG8V": ["Afficher les fichiers (", ["sidebarShortcutLabel"], ")"],
@@ -4100,9 +4099,6 @@
"Colore et marque les fichiers qui ont des problèmes de vérification ou de liens, sans les ouvrir."
],
"ku0yLQ": ["Les pages reçoivent leur embedding la première fois qu'une recherche en a besoin."],
- "kwLcuN": [
- "Donnez aux outils IA que vous utilisez l'accès en lecture et en écriture à vos projets. Choisissez ci-dessous ce qu'il faut configurer, et modifiez-le à tout moment dans Paramètres > Outils IA et CLI."
- ],
"kwNS82": ["Ce thème ne peut pas être utilisé tant que le fichier n’est pas corrigé."],
"ky9Tuc": ["Modifier la source de ", ["0"]],
"kzpkcn": ["Impossible d'enregistrer cette modification."],
@@ -4828,6 +4824,9 @@
"tHKQ6j": ["Renommage"],
"tKlWWY": ["Émoji"],
"tLX6LY": ["OpenKnowledge a écrit un document"],
+ "tLd2qa": [
+ "Aucun outil IA détecté pour l'instant. Une fois que vous en installez un, connectez-le depuis Paramètres > Outils IA et CLI."
+ ],
"tM5uzl": [
"Les familles de provenance, de confiance et de cycle de vie : sources, generated, verified, status et stale_after. Chaque champ est facultatif et l'absence n'est jamais une erreur."
],
@@ -5168,6 +5167,7 @@
"xEb24l": ["Autorisez-nous à vous recontacter par e-mail"],
"xFQRu4": ["Je l'ai déjà en local"],
"xG4tb6": ["Connexion perdue avec « ", ["docName"], " »."],
+ "xGVfLh": ["Continuer"],
"xHGM4U": ["Masquer l'aperçu HTML"],
"xK291b": ["Deux pages (impaires)"],
"xKDsGS": ["Aller à la ligne ", ["0"], " dans ", ["1"]],
@@ -5270,6 +5270,9 @@
"ycIX5i": ["Vous êtes déconnecté — connectez-vous pour reprendre la synchronisation."],
"yckias": ["Renommer le fichier"],
"ydzS9x": ["Quitter"],
+ "yeMZk8": [
+ "Ajoute la connexion MCP d'OpenKnowledge à chacun, ainsi que la compétence open-knowledge-discovery pour que vos agents reconnaissent les projets OpenKnowledge."
+ ],
"yf5auF": ["Rechercher des règles par id, alias ou nom"],
"yf5r22": [
"Type de ",
diff --git a/packages/app/src/locales/fr/messages.po b/packages/app/src/locales/fr/messages.po
index a6c3df0f5..ca649e295 100644
--- a/packages/app/src/locales/fr/messages.po
+++ b/packages/app/src/locales/fr/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {Aucun problème dans # document.} many {Aucun problème
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {Extension} many {Extensions} other {Extensions}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {Afficher # outil de plus} many {Afficher # outils de plus} other {Afficher # outils de plus}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "Ajoute un bloc géré à"
msgid "Adds a managed block to {0}"
msgstr "Ajoute un bloc géré à {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "Ajoute une entrée MCP OpenKnowledge et la compétence du projet, pour que chaque outil puisse lire et mettre à jour ces documents."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "Ajoute la connexion MCP d'OpenKnowledge à chacun, ainsi que la compétence open-knowledge-discovery pour que vos agents reconnaissent les projets OpenKnowledge."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "Ajoute la connexion MCP d'OpenKnowledge à chacun."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "Ajoute la télémétrie, l'état du serveur et les informations d'exécution lorsqu'elles sont disponibles. Les identifiants sont toujours supprimés ; les noms de documents, s'ils sont inclus, apparaissent en clair (non masqués)."
@@ -1469,9 +1479,12 @@ msgstr "Le panneau des agents a été fermé pour que Terminal reste lisible."
msgid "Agent settings"
msgstr "Paramètres de l'agent"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Compétence d'agent"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Compétences d'agent"
@@ -1488,6 +1501,8 @@ msgstr "Agents"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "La gestion des outils IA n'est disponible que dans l'application de bureau OpenKnowledge."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "Vérification des compétences"
msgid "Checking source"
msgstr "Vérification de la source"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "Vérification des outils IA dont vous disposez"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "Vérification de votre connexion GitHub"
@@ -2925,12 +2944,15 @@ msgstr "Fichiers en conflit : {0}."
msgid "Conflicts"
msgstr "Conflits"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "Connecter"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "Connecter {0} à ce projet"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "Connectez un compte GitHub pour parcourir et synchroniser vos dépôts."
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "Connectez ce projet à GitHub pour le partager."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "Connecter aux outils IA"
@@ -3063,7 +3084,6 @@ msgstr "Règles de contenu"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "Les règles de contenu ne sont pas encore chargées — réessayez dans un instant"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "Coût de contexte"
@@ -3076,6 +3096,11 @@ msgstr "Fenêtre de contexte :"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "Fenêtre de contexte : {percent}% utilisée, {usedLabel} tokens sur {sizeLabel}"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "Continuer"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "Continuer dans le navigateur"
@@ -4661,10 +4686,6 @@ msgstr "par ex. L'éditeur s'est figé après avoir collé un grand tableau"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "Chaque pack équipe votre projet de dossiers et de modèles prêts à l'emploi."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "Chaque outil sélectionné reçoit une entrée MCP OpenKnowledge."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "Saisissez une URL valide, par exemple https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "Saisissez l'origine publique servie par votre tunnel, par exemple https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "Champs"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "Donnez aux outils IA que vous utilisez l'accès en lecture et en écriture à vos projets. Cocher une case configure l'outil immédiatement ; décocher le retire."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "Donnez aux outils IA que vous utilisez l'accès en lecture et en écriture à vos projets. Choisissez ci-dessous ce qu'il faut configurer, et modifiez-le à tout moment dans Paramètres > Outils IA et CLI."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "Donnez aux outils IA que vous utilisez l'accès en lecture et en écriture à vos projets. Vous pourrez tout modifier plus tard dans Paramètres > Outils IA et CLI."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "Masquer les options avancées"
msgid "Hide Agents"
msgstr "Masquer les agents"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "Masquer les détails"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "Comment pousser manuellement"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "Comment configurer"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "Comment configurer {0} (s'ouvre dans le navigateur)"
@@ -6795,7 +6815,6 @@ msgstr "Installation"
msgid "Installing {agentName}…"
msgstr "Installation de {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "S'installe dans"
@@ -7438,7 +7457,7 @@ msgstr "La réparation automatique de MCP a échoué"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "Aucun agent sélectionné — les compétences sont enregistrées mais i
msgid "No AI agents installed"
msgstr "Aucun agent IA installé"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "Aucun outil IA détecté pour l'instant. Une fois que vous en installez un, connectez-le depuis Paramètres > Outils IA et CLI."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "Aucun outil IA détecté pour l'instant. Une fois que vous en installez un, connectez-le depuis Paramètres > Ce projet."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "Aucun outil IA détecté. Installez-en un pour utiliser cette compétence."
@@ -10179,11 +10207,6 @@ msgstr "« {projectName} » retiré des projets récents. Son dossier n'existe p
msgid "Removes from"
msgstr "Supprime de"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "Retire cette compétence de vos éditeurs."
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "Retrait en cours"
@@ -10338,6 +10361,11 @@ msgstr "Remplacer ces copies par des liens symboliques vers la source ?"
msgid "Replace these symlinks with independent copies?"
msgstr "Remplacer ces liens symboliques par des copies indépendantes ?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "Remplace l'entrée OpenKnowledge existante dans {0}"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "Configurez-le sur sa propre page sous Plugins."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "Définissez la synchronisation par défaut pour les utilisateurs qui ouvrent ce projet pour la première fois. Ce réglage est validé dans votre dépôt."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "Configurer {0}"
@@ -11405,9 +11433,7 @@ msgstr "Afficher le panneau du document"
msgid "Show external URL nodes"
msgstr "Afficher les nœuds d'URL externes"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "Afficher moins"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "Les compétences sont maintenant partagées. Validez .ok/skills pour les partager avec votre équipe."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "Les compétences s'installent dans tous les outils IA détectés sur cette machine, indépendamment des connexions MCP choisies ci-dessus."
@@ -11655,11 +11679,6 @@ msgstr "skills.sh est indisponible ; affichage des résultats GitHub (sans class
msgid "Skip"
msgstr "Ignorer"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "Passer pour l'instant"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "Que doit faire l'IA ? (facultatif)"
msgid "What the agent may read, write, and surface."
msgstr "Ce que l'agent peut lire, écrire et faire remonter."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "Ce que cela change"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "Sera créé dans : <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "Sera enregistré sous <0>{sanitized}</0>."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "Remplacera l'entrée OpenKnowledge existante"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/hi/messages.json b/packages/app/src/locales/hi/messages.json
index 93fcbc9bc..e4eb6c96f 100644
--- a/packages/app/src/locales/hi/messages.json
+++ b/packages/app/src/locales/hi/messages.json
@@ -303,7 +303,6 @@
"3pU8e-": ["टैब पिन करें"],
"3qDsdf": ["कौशल स्थापित करें"],
"3ri_4O": ["आपके ज्ञानकोष को Open Knowledge Format के अनुरूप रखता है।"],
- "3s-k7L": [["0", "plural", { "one": ["#", " और टूल दिखाएँ"], "other": ["#", " और टूल दिखाएँ"] }]],
"3uSzeS": ["छवियाँ कहीं से भी · गैर-छवि फ़ाइलें कार्यक्षेत्र में होनी चाहिए"],
"3v62tp": ["skills.sh पर लोकप्रिय"],
"3wiBAs": ["यह कहाँ रहना चाहिए?"],
@@ -604,7 +603,6 @@
["0"],
" पर रहती है — वह फ़ोल्डर यह स्किल नहीं है। सुलझाने के लिए दोनों में से एक का नाम बदलें या हटाएँ।"
],
- "8IP5jY": ["हर चयनित टूल को एक OpenKnowledge MCP प्रविष्टि मिलती है।"],
"8Jm2EP": ["शीर्षक स्तर 1 से 6 तक चालू या बंद करें।"],
"8L0EPw": ["या कोई टोकन चिपकाएँ"],
"8OGgNM": ["दृश्य-संपादक की खोज खोलें या बंद करें।"],
@@ -851,6 +849,9 @@
"C0cll5": [["0"], " पर इंस्टॉल है — हटाने के लिए क्लिक करें"],
"C0hoX6": ["हमेशा शामिल"],
"C3Zc44": ["स्किल खोजें"],
+ "C4HK3A": [
+ "OpenKnowledge की MCP प्रविष्टि और प्रोजेक्ट Skill जोड़ता है, ताकि हर टूल इन दस्तावेज़ों को पढ़ और अपडेट कर सके।"
+ ],
"C62P6b": ["संपादक चयन"],
"C7PKTv": ["चुनी हुई टिप्पणियाँ चैट में भेजें"],
"C7waqI": [["0", "plural", { "one": ["#", " दस्तावेज़"], "other": ["#", " दस्तावेज़"] }]],
@@ -1472,6 +1473,7 @@
"L83DKQ": ["PATH इंस्टॉल विफल रहा: ", ["summary"]],
"L8fKuc": ["इसे अपने टर्मिनल में चलाएँ"],
"L9Y9Om": ["सबसे ज़्यादा लिंक किए गए पेज"],
+ "LAR9ub": [["0"], " में मौजूदा OpenKnowledge प्रविष्टि को बदल देगा"],
"LBv_TR": ["स्थानों को सिमलिंक में बदलें"],
"LBwUR6": ["ब्रांच बदलने में समय समाप्त — ", ["targetNoun"], " को हाथ से खोलने की कोशिश करें।"],
"LCUy_N": ["सत्र समाप्त"],
@@ -1717,6 +1719,9 @@
"ODGGlj": ["फ़ोल्डर का नाम: ", ["detail"]],
"OEEQKT": ["अच्छा"],
"OETysD": ["अभी कोई हब पेज नहीं। पेजों पर आने वाले ग्राफ़ लिंक जमा होने के बाद हब दिखते हैं।"],
+ "OFIL46": [
+ "आप जिन AI टूल का उपयोग करते हैं उन्हें अपने प्रोजेक्ट पढ़ने और अपडेट करने की अनुमति दें। यह सब बाद में Settings > AI tools & CLI में बदला जा सकता है।"
+ ],
"OJ4VDl": [
"यह एजेंट ",
["0"],
@@ -2057,7 +2062,6 @@
"T3j-7o": ["जैसे सिंक चलते समय प्रोजेक्ट बदलना"],
"TBfy1C": ["Slidev शुरू नहीं किया जा सका।"],
"TDeOpa": ["अपने ही संपादन पूर्ववत या फिर से करें, सहयोगियों और एजेंट के संपादन अपनी जगह छोड़ते हुए।"],
- "TEEPPF": ["मौजूदा OpenKnowledge प्रविष्टि की जगह ले लेगा"],
"TEJiKw": ["OpenKnowledge ने ", ["subject"], " की जाँच की"],
"TKQ7K-": ["इंस्टॉल करें"],
"TLIO3j": ["तिरछी फ़ॉर्मैटिंग चालू या बंद करें।"],
@@ -2647,6 +2651,7 @@
"। इनमें से कोई भी प्लगइन से इंस्टॉल करें।"
],
"aRI3Om": ["विंडो बंद करें"],
+ "aSVmqV": ["अभी तक कोई AI टूल नहीं मिला। एक इंस्टॉल करने के बाद उसे Settings > This project से जोड़ें।"],
"aWlJai": ["साइन-इन प्रक्रिया शुरू की जा रही है"],
"aXkLuE": ["अप्लाई प्रतिक्रिया का प्रारूप गलत है"],
"a_2_uB": [["trimmed"], " नहीं बनाया जा सका: ", ["0"]],
@@ -2845,7 +2850,6 @@
"dVaXnv": ["सामग्री नियम अभी लोड नहीं हुए — कुछ देर बाद फिर कोशिश करें"],
"dXCi3J": ["कुछ टूट गया"],
"dXOolp": ["आपके बदलाव सिंक हो रहे हैं"],
- "dXWMFn": ["इस स्किल को आपके संपादकों से हटाता है।"],
"darW93": [
[
"pageCount",
@@ -2898,6 +2902,7 @@
"ePK91l": ["संपादित करें"],
"eQkgKV": ["इंस्टॉल है"],
"eQn7nI": ["कॉन्फ़िग साझाकरण लोड हो रहा है"],
+ "eQsWFc": ["Agent Skill"],
"eQxm13": ["संलग्नक स्थान अपडेट करने में विफल — ", ["detail"]],
"eRSPr9": ["टर्मिनल खोलें"],
"eTL5y9": ["चयन पूर्वावलोकन दिखाएँ"],
@@ -2977,6 +2982,7 @@
"fhaxJL": ["क्लोन विफल — प्रतिक्रिया का मुख्य भाग खाली है"],
"fiMNLl": ["सिंक सेटिंग अभी लोड नहीं हुई — कुछ देर बाद फिर कोशिश करें"],
"fj5YiP": ["GitHub से एक नई प्रति डाउनलोड करता है।"],
+ "fmwwju": ["हर टूल में OpenKnowledge का MCP कनेक्शन जोड़ता है।"],
"fnlBN4": ["स्वत: स्वीकृति सेटिंग अभी लोड नहीं हुई — कुछ देर बाद फिर कोशिश करें"],
"fo0VXg": ["अनइंस्टॉल करें"],
"fo6CCk": ["फलक ", ["0"]],
@@ -3131,7 +3137,6 @@
"hmYd9H": [
"आप जिन लोगों और कंपनियों के साथ काम करते हैं और उनके साथ होने वाली बैठकों पर नज़र रखें। यह याद रखने के लिए अच्छा कि कोई कौन है, आप उसे कैसे जानते हैं, और आपकी आख़िरी बात किस पर हुई थी।"
],
- "hnlGzG": ["अभी के लिए छोड़ें"],
"hnq8fP": [["label"], " को संदर्भ से हटाएँ"],
"hp8OtS": ["जोड़ा गया"],
"hpDtUm": ["बाहरी विकी लिंक"],
@@ -3248,6 +3253,7 @@
"j2kt1j": [
"अपनी पसंदीदा AI के साथ एक नया सत्र खोलें। पाठ चयनित होने पर, उसे चलाने के बजाय वहीं तैयार रखें।"
],
+ "j4SmVY": [["0"], " को इस प्रोजेक्ट से जोड़ें"],
"j5CWO4": ["अद्यतित"],
"j9Vqvc": ["विरासत में मिला"],
"j9tqwO": ["साइडबार में नहीं"],
@@ -3274,6 +3280,7 @@
],
"jSY41h": ["\"", ["docName"], "\" में बहुत समय लगा। अपना कनेक्शन जाँचें।"],
"jTLWxh": ["एक बाहरी पेज को इनलाइन iframe में एम्बेड करें (दस्तावेज़, डेमो, Figma, CodeSandbox)।"],
+ "jWzDfC": ["जाँच रहे हैं कि आपके पास कौन-से AI टूल हैं"],
"jZ_E5M": ["फ़ेच विफल — विवरण के लिए सर्वर लॉग देखें।"],
"j_Z6g7": ["साझा डिफ़ॉल्ट"],
"jbSG8V": ["Files दिखाएँ (", ["sidebarShortcutLabel"], ")"],
@@ -3402,9 +3409,6 @@
"kpqWwG": ["यह ", ["targetNoun"], " ब्रांच <0>", ["branch"], "</0> से हटा दिया गया था।"],
"ktspdq": ["जिन फ़ाइलों में lint या लिंक की समस्याएँ हैं उन्हें बिना खोले रंग और बैज दें।"],
"ku0yLQ": ["पेज पहली बार तभी एम्बेड होते हैं जब किसी खोज को उनकी ज़रूरत पड़ती है।"],
- "kwLcuN": [
- "आप जिन AI टूल का उपयोग करते हैं उन्हें अपने प्रोजेक्ट पढ़ने और अपडेट करने की पहुँच दें। नीचे चुनें कि क्या सेट अप करना है, और इसे कभी भी Settings > AI tools & CLI में बदलें।"
- ],
"kwNS82": ["फ़ाइल ठीक होने तक इस थीम का उपयोग नहीं किया जा सकता।"],
"ky9Tuc": [["0"], " स्रोत संपादित करें"],
"kzpkcn": ["वह संपादन नहीं सहेजा जा सका।"],
@@ -4000,6 +4004,7 @@
"tHKQ6j": ["नाम बदला जा रहा है"],
"tKlWWY": ["इमोजी"],
"tLX6LY": ["OpenKnowledge ने एक दस्तावेज़ लिखा"],
+ "tLd2qa": ["अभी तक कोई AI टूल नहीं मिला। एक इंस्टॉल करने के बाद उसे Settings > AI tools & CLI से जोड़ें।"],
"tM5uzl": [
"उत्पत्ति, विश्वास और जीवनचक्र परिवार: sources, generated, verified, status, और stale_after। प्रत्येक क्षेत्र वैकल्पिक है, और अनुपस्थिति कभी भी कोई त्रुटि नहीं होती है।"
],
@@ -4278,6 +4283,7 @@
"xEb24l": ["हमें ईमेल से आगे संपर्क करने दें"],
"xFQRu4": ["यह मेरे पास पहले से स्थानीय रूप से है"],
"xG4tb6": ["\"", ["docName"], "\" से कनेक्शन टूट गया।"],
+ "xGVfLh": ["जारी रखें"],
"xHGM4U": ["HTML पूर्वावलोकन छिपाएँ"],
"xK291b": ["दो पेज (विषम)"],
"xKDsGS": [["1"], " में पंक्ति ", ["0"], " पर जाएँ"],
@@ -4364,6 +4370,9 @@
"ycIX5i": ["आप साइन आउट हैं — सिंक फिर से शुरू करने के लिए साइन इन करें।"],
"yckias": ["फ़ाइल का नाम बदलें"],
"ydzS9x": ["बाहर निकलें"],
+ "yeMZk8": [
+ "हर टूल में OpenKnowledge का MCP कनेक्शन जोड़ता है, साथ ही open-knowledge-discovery Skill भी, ताकि आपके एजेंट OpenKnowledge प्रोजेक्ट पहचान सकें।"
+ ],
"yf5auF": ["नियम id, उपनाम या नाम से खोजें"],
"yf5r22": [
["keyName"],
diff --git a/packages/app/src/locales/hi/messages.po b/packages/app/src/locales/hi/messages.po
index b1d6fa6e0..4d517cbaa 100644
--- a/packages/app/src/locales/hi/messages.po
+++ b/packages/app/src/locales/hi/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {# दस्तावेज़ में कोई स
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {प्लगइन} other {प्लगइन}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {# और टूल दिखाएँ} other {# और टूल दिखाएँ}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1410,6 +1405,21 @@ msgstr "इसमें एक प्रबंधित ब्लॉक जो
msgid "Adds a managed block to {0}"
msgstr "{0} में एक प्रबंधित ब्लॉक जोड़ता है"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "OpenKnowledge की MCP प्रविष्टि और प्रोजेक्ट Skill जोड़ता है, ताकि हर टूल इन दस्तावेज़ों को पढ़ और अपडेट कर सके।"
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "हर टूल में OpenKnowledge का MCP कनेक्शन जोड़ता है, साथ ही open-knowledge-discovery Skill भी, ताकि आपके एजेंट OpenKnowledge प्रोजेक्ट पहचान सकें।"
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "हर टूल में OpenKnowledge का MCP कनेक्शन जोड़ता है।"
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "उपलब्ध होने पर टेलीमेट्री, सर्वर स्थिति और रनटाइम जानकारी जोड़ता है। क्रेडेंशियल हमेशा हटा दिए जाते हैं; दस्तावेज़ों के नाम, यदि शामिल हों, सादे पाठ में दिखते हैं (हटाए नहीं जाते)।"
@@ -1465,9 +1475,12 @@ msgstr "Terminal को पढ़ने योग्य बनाए रखन
msgid "Agent settings"
msgstr "एजेंट सेटिंग"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Agent Skill"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Agent Skills"
@@ -1484,6 +1497,8 @@ msgstr "एजेंट"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI टूल प्रबंधन केवल OpenKnowledge डेस्कटॉप ऐप में उपलब्ध है।"
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2371,6 +2386,10 @@ msgstr "कौशल की जाँच करना"
msgid "Checking source"
msgstr "स्रोत जाँचा जा रहा है"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "जाँच रहे हैं कि आपके पास कौन-से AI टूल हैं"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "आपका GitHub कनेक्शन जाँचा जा रहा है"
@@ -2921,12 +2940,15 @@ msgstr "विरोधी फ़ाइलें: {0}।"
msgid "Conflicts"
msgstr "विरोध"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "जोड़ें"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "{0} को इस प्रोजेक्ट से जोड़ें"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "अपनी रिपॉज़िटरी ब्राउज़ और सिंक करने के लिए कोई GitHub खाता जोड़ें।"
@@ -2960,7 +2982,6 @@ msgid "Connect this project to GitHub to share."
msgstr "साझा करने के लिए इस प्रोजेक्ट को GitHub से जोड़ें।"
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "AI टूल से जोड़ें"
@@ -3059,7 +3080,6 @@ msgstr "सामग्री नियम"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "सामग्री नियम अभी लोड नहीं हुए — कुछ देर बाद फिर कोशिश करें"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "कॉन्टेक्स्ट लागत"
@@ -3072,6 +3092,11 @@ msgstr "संदर्भ विंडो:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "संदर्भ विंडो: {percent}% उपयोग में, {sizeLabel} में से {usedLabel} टोकन"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "जारी रखें"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "ब्राउज़र में जारी रखें"
@@ -4657,10 +4682,6 @@ msgstr "जैसे एक बड़ी तालिका चिपकान
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "हर पैक आपके प्रोजेक्ट को तैयार फ़ोल्डर और टेम्पलेट के साथ ढाँचा देता है।"
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "हर चयनित टूल को एक OpenKnowledge MCP प्रविष्टि मिलती है।"
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5017,7 +5038,6 @@ msgstr "एक मान्य URL दर्ज करें, उदाहरण
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "वह सार्वजनिक ऑरिजिन दर्ज करें जिसे आपका tunnel सर्व करता है, उदाहरण के लिए https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5527,7 +5547,6 @@ msgstr "फ़ील्ड"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5977,8 +5996,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "आप जिन AI टूल का उपयोग करते हैं उन्हें अपने प्रोजेक्ट पढ़ने और अपडेट करने की पहुँच दें। बॉक्स चुनते ही यह तुरंत सेट हो जाता है; हटाने पर हट जाता है।"
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "आप जिन AI टूल का उपयोग करते हैं उन्हें अपने प्रोजेक्ट पढ़ने और अपडेट करने की पहुँच दें। नीचे चुनें कि क्या सेट अप करना है, और इसे कभी भी Settings > AI tools & CLI में बदलें।"
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "आप जिन AI टूल का उपयोग करते हैं उन्हें अपने प्रोजेक्ट पढ़ने और अपडेट करने की अनुमति दें। यह सब बाद में Settings > AI tools & CLI में बदला जा सकता है।"
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6160,7 +6179,11 @@ msgstr "उन्नत छिपाएँ"
msgid "Hide Agents"
msgstr "Agents छिपाएँ"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "विवरण छिपाएँ"
@@ -6302,14 +6325,11 @@ msgid "How to push manually"
msgstr "हाथ से कैसे भेजें"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "सेट अप कैसे करें"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "{0} कैसे सेट अप करें (ब्राउज़र में खुलता है)"
@@ -6791,7 +6811,6 @@ msgstr "इंस्टॉल किया जा रहा है"
msgid "Installing {agentName}…"
msgstr "{agentName} इंस्टॉल किया जा रहा है…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "यहाँ इंस्टॉल होगा"
@@ -7434,7 +7453,7 @@ msgstr "MCP स्वतः-मरम्मत विफल रही"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8042,6 +8061,15 @@ msgstr "कोई एजेंट नहीं चुना गया — स्
msgid "No AI agents installed"
msgstr "कोई AI एजेंट इंस्टॉल नहीं है"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "अभी तक कोई AI टूल नहीं मिला। एक इंस्टॉल करने के बाद उसे Settings > AI tools & CLI से जोड़ें।"
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "अभी तक कोई AI टूल नहीं मिला। एक इंस्टॉल करने के बाद उसे Settings > This project से जोड़ें।"
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "कोई AI टूल नहीं मिला। इस स्किल को इस्तेमाल करने के लिए एक इंस्टॉल करें।"
@@ -10175,11 +10203,6 @@ msgstr "\"{projectName}\" को हाल के प्रोजेक्ट
msgid "Removes from"
msgstr "यहाँ से हटेगा"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "इस स्किल को आपके संपादकों से हटाता है।"
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "हटाया जा रहा है"
@@ -10334,6 +10357,11 @@ msgstr "इन प्रतियों को स्रोत के सिम
msgid "Replace these symlinks with independent copies?"
msgstr "इन सिमलिंक को स्वतंत्र प्रतियों से बदलें?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "{0} में मौजूदा OpenKnowledge प्रविष्टि को बदल देगा"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11209,7 +11237,7 @@ msgstr "इसे Plugins के अंतर्गत इसके अपने
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "इस प्रोजेक्ट को पहली बार खोलने वाले उपयोगकर्ताओं के लिए सिंक का डिफ़ॉल्ट तय करें। यह सेटिंग आपकी रिपॉज़िटरी में कमिट होती है।"
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "{0} सेट अप करें"
@@ -11401,9 +11429,7 @@ msgstr "दस्तावेज़ पैनल दिखाएँ"
msgid "Show external URL nodes"
msgstr "बाहरी URL नोड दिखाएँ"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "कम दिखाएँ"
@@ -11622,8 +11648,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "स्किल अब साझा हैं। अपनी टीम के साथ साझा करने के लिए .ok/skills कमिट करें।"
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "स्किल इस मशीन पर मिले हर AI टूल में इंस्टॉल होती हैं, चाहे आपने ऊपर कोई भी MCP कनेक्शन चुना हो।"
@@ -11651,11 +11675,6 @@ msgstr "skills.sh अनुपलब्ध है; GitHub के परिणा
msgid "Skip"
msgstr "छोड़ें"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "अभी के लिए छोड़ें"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13989,11 +14008,13 @@ msgstr "AI को क्या करना चाहिए? (वैकल्प
msgid "What the agent may read, write, and surface."
msgstr "एजेंट क्या पढ़, लिख और सामने ला सकता है।"
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "यह क्या बदलता है"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14087,11 +14108,6 @@ msgstr "यहाँ बनाया जाएगा: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "<0>{sanitized}</0> के रूप में सहेजा जाएगा।"
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "मौजूदा OpenKnowledge प्रविष्टि की जगह ले लेगा"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/id/messages.json b/packages/app/src/locales/id/messages.json
index daed8ebcd..7753371ff 100644
--- a/packages/app/src/locales/id/messages.json
+++ b/packages/app/src/locales/id/messages.json
@@ -307,7 +307,6 @@
"3pU8e-": ["Sematkan tab"],
"3qDsdf": ["Instal keterampilan"],
"3ri_4O": ["Menjaga basis pengetahuan Anda tetap selaras dengan Open Knowledge Format."],
- "3s-k7L": [["0", "plural", { "other": ["Tampilkan ", "#", " alat lagi"] }]],
"3uSzeS": ["Gambar dari mana saja · berkas non-gambar harus berada di ruang kerja"],
"3v62tp": ["Populer di skills.sh"],
"3wiBAs": ["Di mana ia sebaiknya berada?"],
@@ -624,7 +623,6 @@
["0"],
" — folder itu bukan keterampilan ini. Ganti nama atau hapus salah satu dari keduanya untuk menyelesaikannya."
],
- "8IP5jY": ["Setiap alat yang dipilih mendapat entri MCP OpenKnowledge."],
"8Jm2EP": ["Menyalakan atau mematikan tingkat judul 1 sampai 6."],
"8L0EPw": ["atau tempelkan sebuah token"],
"8OGgNM": ["Membuka atau menutup pencarian penyunting visual."],
@@ -864,6 +862,9 @@
"C0cll5": ["Terpasang di ", ["0"], " — klik untuk menghapus"],
"C0hoX6": ["Selalu disertakan"],
"C3Zc44": ["Jelajahi keterampilan"],
+ "C4HK3A": [
+ "Menambahkan entri MCP OpenKnowledge dan keterampilan proyek, sehingga setiap alat bisa membaca dan memperbarui dokumen ini."
+ ],
"C62P6b": ["Pilihan penyunting"],
"C7PKTv": ["Kirim komentar yang dicentang ke obrolan"],
"C7waqI": [["0", "plural", { "other": ["#", " dokumen"] }]],
@@ -1528,6 +1529,7 @@
"L83DKQ": ["Pemasangan PATH gagal: ", ["summary"]],
"L8fKuc": ["Jalankan ini di terminal Anda"],
"L9Y9Om": ["Halaman paling banyak ditaut"],
+ "LAR9ub": ["Mengganti entri OpenKnowledge yang ada di ", ["0"]],
"LBv_TR": ["Ubah lokasi menjadi symlink"],
"LBwUR6": [
"Perpindahan cabang kehabisan waktu — coba buka ",
@@ -1774,6 +1776,9 @@
"OETysD": [
"Belum ada halaman pusat. Pusat muncul begitu halaman mengumpulkan tautan graf masuk."
],
+ "OFIL46": [
+ "Beri alat AI yang Anda pakai akses untuk membaca dan memperbarui proyek Anda. Anda bisa mengubah semuanya nanti di Pengaturan > Alat AI & CLI."
+ ],
"OJ4VDl": [
"Agen ini berjalan lewat ",
["0"],
@@ -2122,7 +2127,6 @@
"TDeOpa": [
"Membatalkan atau mengulangi suntingan Anda sendiri, tanpa mengusik suntingan kolaborator dan agen."
],
- "TEEPPF": ["Akan menggantikan entri OpenKnowledge yang sudah ada"],
"TEJiKw": ["OpenKnowledge memeriksa ", ["subject"]],
"TKQ7K-": ["Pasang"],
"TLIO3j": ["Menyalakan atau mematikan format miring."],
@@ -2728,6 +2732,9 @@
". Pasang salah satunya dari plugin."
],
"aRI3Om": ["Tutup Jendela"],
+ "aSVmqV": [
+ "Belum ada alat AI yang terdeteksi. Setelah Anda memasang satu, hubungkan dari Pengaturan > Proyek ini."
+ ],
"aWlJai": ["Memulai alur masuk"],
"aXkLuE": ["Respons apply salah bentuk"],
"a_2_uB": [["trimmed"], " tidak bisa dibuat: ", ["0"]],
@@ -2947,7 +2954,6 @@
"dVaXnv": ["Aturan konten belum dimuat — coba lagi sebentar"],
"dXCi3J": ["Ada yang rusak"],
"dXOolp": ["Menyinkronkan perubahan Anda"],
- "dXWMFn": ["Menghapus keterampilan ini dari penyunting Anda."],
"darW93": [
[
"pageCount",
@@ -3005,6 +3011,7 @@
"ePK91l": ["Sunting"],
"eQkgKV": ["Terpasang"],
"eQn7nI": ["Memuat berbagi konfigurasi"],
+ "eQsWFc": ["Keterampilan Agen"],
"eQxm13": ["Gagal memperbarui lokasi lampiran — ", ["detail"]],
"eRSPr9": ["Buka terminal"],
"eTL5y9": ["Tampilkan pratinjau pilihan"],
@@ -3088,6 +3095,7 @@
"fhaxJL": ["Kloning gagal — badan respons kosong"],
"fiMNLl": ["Pengaturan sinkronisasi belum dimuat — coba lagi sebentar"],
"fj5YiP": ["Mengunduh salinan baru dari GitHub."],
+ "fmwwju": ["Menambahkan koneksi MCP OpenKnowledge ke masing-masing."],
"fnlBN4": ["Pengaturan persetujuan otomatis belum dimuat — coba lagi sebentar"],
"fo0VXg": ["Copot"],
"fo6CCk": ["panel ", ["0"]],
@@ -3257,7 +3265,6 @@
"hmYd9H": [
"Catat orang dan perusahaan yang bekerja dengan Anda serta pertemuan yang Anda adakan dengan mereka. Cocok untuk mengingat siapa seseorang, dari mana Anda mengenalnya, dan apa yang terakhir Anda bicarakan."
],
- "hnlGzG": ["Lewati dulu"],
"hnq8fP": ["Hapus ", ["label"], " dari konteks"],
"hp8OtS": ["Ditambahkan"],
"hpDtUm": ["Tautan wiki eksternal"],
@@ -3385,6 +3392,7 @@
"j2kt1j": [
"Membuka sesi baru dengan AI pilihan Anda. Dengan teks terpilih, teks itu disiapkan di sana alih-alih dijalankan."
],
+ "j4SmVY": ["Hubungkan ", ["0"], " ke proyek ini"],
"j5CWO4": ["Sudah mutakhir"],
"j9Vqvc": ["diwarisi"],
"j9tqwO": ["Tidak ada di bilah sisi"],
@@ -3413,6 +3421,7 @@
"jTLWxh": [
"Menyematkan halaman eksternal dalam iframe sebaris (dokumentasi, demo, Figma, CodeSandbox)."
],
+ "jWzDfC": ["Memeriksa alat AI yang Anda miliki"],
"jZ_E5M": ["Fetch gagal — periksa log server untuk detailnya."],
"j_Z6g7": ["Bawaan bersama"],
"jbSG8V": ["Tampilkan Berkas (", ["sidebarShortcutLabel"], ")"],
@@ -3543,9 +3552,6 @@
"Mewarnai dan menandai berkas yang punya masalah lint atau tautan, tanpa perlu membukanya."
],
"ku0yLQ": ["Halaman di-embed pertama kali ketika sebuah pencarian membutuhkannya."],
- "kwLcuN": [
- "Beri alat AI yang Anda pakai akses untuk membaca dan memperbarui proyek Anda. Pilih apa yang disiapkan di bawah, dan ubah kapan saja di Pengaturan > Alat AI & CLI."
- ],
"kwNS82": ["Tema ini tidak dapat digunakan sampai berkasnya diperbaiki."],
"ky9Tuc": ["Sunting sumber ", ["0"]],
"kzpkcn": ["Suntingan itu tidak bisa disimpan."],
@@ -4163,6 +4169,9 @@
"tHKQ6j": ["Mengganti nama"],
"tKlWWY": ["Emoji"],
"tLX6LY": ["OpenKnowledge menulis sebuah dokumen"],
+ "tLd2qa": [
+ "Belum ada alat AI yang terdeteksi. Setelah Anda memasang satu, hubungkan dari Pengaturan > Alat AI & CLI."
+ ],
"tM5uzl": [
"Keluarga asal, kepercayaan, dan siklus hidup: sources, generated, verified, status, dan stale_after. Setiap bidang bersifat opsional, dan ketidakhadiran tidak pernah merupakan kesalahan."
],
@@ -4449,6 +4458,7 @@
"xEb24l": ["Izinkan kami menindaklanjuti lewat surel"],
"xFQRu4": ["Saya sudah punya secara lokal"],
"xG4tb6": ["Koneksi ke \"", ["docName"], "\" terputus."],
+ "xGVfLh": ["Lanjutkan"],
"xHGM4U": ["Sembunyikan pratinjau HTML"],
"xK291b": ["Dua halaman (ganjil)"],
"xKDsGS": ["Ke baris ", ["0"], " di ", ["1"]],
@@ -4535,6 +4545,9 @@
"ycIX5i": ["Anda sudah keluar — masuk untuk melanjutkan sinkronisasi."],
"yckias": ["Ganti nama berkas"],
"ydzS9x": ["Keluar"],
+ "yeMZk8": [
+ "Menambahkan koneksi MCP OpenKnowledge ke masing-masing, ditambah keterampilan open-knowledge-discovery agar agen Anda mengenali proyek OpenKnowledge."
+ ],
"yf5auF": ["Cari aturan berdasarkan id, alias, atau nama"],
"yf5r22": [
"Tipe ",
diff --git a/packages/app/src/locales/id/messages.po b/packages/app/src/locales/id/messages.po
index 02bce5da2..669d86c8a 100644
--- a/packages/app/src/locales/id/messages.po
+++ b/packages/app/src/locales/id/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, other {Tidak ada masalah di # dokumen.}}"
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, other {Plugin}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, other {Tampilkan # alat lagi}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1410,6 +1405,21 @@ msgstr "Menambahkan blok terkelola ke"
msgid "Adds a managed block to {0}"
msgstr "Menambahkan blok terkelola ke {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "Menambahkan entri MCP OpenKnowledge dan keterampilan proyek, sehingga setiap alat bisa membaca dan memperbarui dokumen ini."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "Menambahkan koneksi MCP OpenKnowledge ke masing-masing, ditambah keterampilan open-knowledge-discovery agar agen Anda mengenali proyek OpenKnowledge."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "Menambahkan koneksi MCP OpenKnowledge ke masing-masing."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "Menambahkan telemetri, keadaan server, dan info runtime bila tersedia. Kredensial selalu dihapus; nama dokumen, jika disertakan, muncul sebagai teks biasa (tidak disunting)."
@@ -1465,9 +1475,12 @@ msgstr "Panel agen ditutup agar Terminal tetap mudah dibaca."
msgid "Agent settings"
msgstr "Pengaturan agen"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Keterampilan Agen"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Keterampilan Agen"
@@ -1484,6 +1497,8 @@ msgstr "Agen"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "Pengelolaan alat AI hanya tersedia di aplikasi desktop OpenKnowledge."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2371,6 +2386,10 @@ msgstr "Memeriksa keterampilan"
msgid "Checking source"
msgstr "Memeriksa sumber"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "Memeriksa alat AI yang Anda miliki"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "Memeriksa koneksi GitHub Anda"
@@ -2921,12 +2940,15 @@ msgstr "Berkas yang berkonflik: {0}."
msgid "Conflicts"
msgstr "Konflik"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "Hubungkan"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "Hubungkan {0} ke proyek ini"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "Hubungkan akun GitHub untuk menelusuri dan menyinkronkan repositori Anda."
@@ -2960,7 +2982,6 @@ msgid "Connect this project to GitHub to share."
msgstr "Hubungkan proyek ini ke GitHub untuk membagikannya."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "Hubungkan ke alat AI"
@@ -3059,7 +3080,6 @@ msgstr "Aturan konten"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "Aturan konten belum dimuat — coba lagi sebentar"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "Biaya konteks"
@@ -3072,6 +3092,11 @@ msgstr "Jendela konteks:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "Jendela konteks: {percent}% terpakai, {usedLabel} dari {sizeLabel} token"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "Lanjutkan"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "Lanjutkan di peramban"
@@ -4657,10 +4682,6 @@ msgstr "mis. Penyunting membeku setelah saya menempelkan tabel besar"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "Setiap paket menyusun proyek Anda dengan folder dan templat siap pakai."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "Setiap alat yang dipilih mendapat entri MCP OpenKnowledge."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5017,7 +5038,6 @@ msgstr "Masukkan URL yang valid, misalnya https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "Masukkan origin publik yang dilayani tunnel Anda, misalnya https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5527,7 +5547,6 @@ msgstr "Bidang"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5977,8 +5996,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "Beri alat AI yang Anda pakai akses untuk membaca dan memperbarui proyek Anda. Mencentang kotak langsung menyiapkannya; melepas centang menghapusnya."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "Beri alat AI yang Anda pakai akses untuk membaca dan memperbarui proyek Anda. Pilih apa yang disiapkan di bawah, dan ubah kapan saja di Pengaturan > Alat AI & CLI."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "Beri alat AI yang Anda pakai akses untuk membaca dan memperbarui proyek Anda. Anda bisa mengubah semuanya nanti di Pengaturan > Alat AI & CLI."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6160,7 +6179,11 @@ msgstr "Sembunyikan opsi lanjutan"
msgid "Hide Agents"
msgstr "Sembunyikan Agen"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "Sembunyikan detail"
@@ -6302,14 +6325,11 @@ msgid "How to push manually"
msgstr "Cara push manual"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "Cara menyiapkan"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "Cara menyiapkan {0} (terbuka di peramban)"
@@ -6791,7 +6811,6 @@ msgstr "Memasang"
msgid "Installing {agentName}…"
msgstr "Memasang {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "Dipasang di"
@@ -7434,7 +7453,7 @@ msgstr "Perbaikan otomatis MCP gagal"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8042,6 +8061,15 @@ msgstr "Tidak ada agen yang dipilih — keterampilannya tersimpan tetapi tidak d
msgid "No AI agents installed"
msgstr "Tidak ada agen AI yang terpasang"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "Belum ada alat AI yang terdeteksi. Setelah Anda memasang satu, hubungkan dari Pengaturan > Alat AI & CLI."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "Belum ada alat AI yang terdeteksi. Setelah Anda memasang satu, hubungkan dari Pengaturan > Proyek ini."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "Tidak ada alat AI yang terdeteksi. Pasang satu untuk memakai keterampilan ini."
@@ -10175,11 +10203,6 @@ msgstr "\"{projectName}\" dihapus dari proyek terbaru. Foldernya sudah tidak ada
msgid "Removes from"
msgstr "Dihapus dari"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "Menghapus keterampilan ini dari penyunting Anda."
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "Menghapus"
@@ -10334,6 +10357,11 @@ msgstr "Ganti salinan-salinan ini dengan symlink ke sumbernya?"
msgid "Replace these symlinks with independent copies?"
msgstr "Ganti symlink-symlink ini dengan salinan mandiri?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "Mengganti entri OpenKnowledge yang ada di {0}"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11209,7 +11237,7 @@ msgstr "Siapkan di halamannya sendiri di bawah Plugin."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "Setel sinkronisasi bawaan untuk pengguna yang membuka proyek ini pertama kali. Pengaturan ini disimpan di repositori Anda."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "Siapkan {0}"
@@ -11401,9 +11429,7 @@ msgstr "Tampilkan panel dokumen"
msgid "Show external URL nodes"
msgstr "Tampilkan simpul URL eksternal"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "Tampilkan lebih sedikit"
@@ -11622,8 +11648,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "Keterampilan kini dibagikan. Lakukan commit pada .ok/skills untuk membagikannya dengan tim Anda."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "Keterampilan dipasang ke setiap alat AI yang terdeteksi di mesin ini, terlepas dari koneksi MCP yang Anda pilih di atas."
@@ -11651,11 +11675,6 @@ msgstr "skills.sh tidak tersedia; menampilkan hasil GitHub (tanpa peringkat pema
msgid "Skip"
msgstr "Lewati"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "Lewati dulu"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13989,11 +14008,13 @@ msgstr "Apa yang harus dilakukan AI? (opsional)"
msgid "What the agent may read, write, and surface."
msgstr "Apa yang boleh dibaca, ditulis, dan ditampilkan agen."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "Apa yang berubah karena ini"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14087,11 +14108,6 @@ msgstr "Akan dibuat di: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "Akan disimpan sebagai <0>{sanitized}</0>."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "Akan menggantikan entri OpenKnowledge yang sudah ada"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/ko/messages.json b/packages/app/src/locales/ko/messages.json
index 052b12136..4e736ef14 100644
--- a/packages/app/src/locales/ko/messages.json
+++ b/packages/app/src/locales/ko/messages.json
@@ -296,7 +296,6 @@
"3pU8e-": ["탭 고정"],
"3qDsdf": ["스킬 설치"],
"3ri_4O": ["지식 베이스를 Open Knowledge Format에 맞게 유지합니다."],
- "3s-k7L": [["0", "plural", { "other": ["도구 ", "#", "개 더 보기"] }]],
"3uSzeS": ["어디서든 가져온 이미지 · 이미지 외 파일, 워크스페이스 필요"],
"3v62tp": ["skills.sh에서 인기"],
"3wiBAs": ["어디에 둘까요?"],
@@ -602,7 +601,6 @@
["0"],
"에 있습니다. 해당 폴더는 이 스킬이 아닙니다. 둘 중 하나를 이름 변경하거나 삭제해 해결하세요."
],
- "8IP5jY": ["선택한 각 도구에 OpenKnowledge MCP 항목이 추가됩니다."],
"8Jm2EP": ["제목 수준 1~6을 전환합니다."],
"8L0EPw": ["또는 토큰 붙여넣기"],
"8OGgNM": ["비주얼 편집기 찾기를 열거나 닫습니다."],
@@ -835,6 +833,9 @@
"C0cll5": [["0"], "에 설치됨. 제거하려면 클릭"],
"C0hoX6": ["항상 포함"],
"C3Zc44": ["스킬 탐색"],
+ "C4HK3A": [
+ "OpenKnowledge MCP 항목과 프로젝트 스킬을 추가하여 각 도구가 이 문서를 읽고 업데이트할 수 있게 합니다."
+ ],
"C62P6b": ["편집기 선택 영역"],
"C7PKTv": ["확인한 댓글을 채팅으로 보내기"],
"C7waqI": [["0", "plural", { "other": ["#", "개 문서"] }]],
@@ -1465,6 +1466,7 @@
"L83DKQ": ["PATH 설치 실패: ", ["summary"]],
"L8fKuc": ["터미널에서 실행"],
"L9Y9Om": ["링크가 가장 많은 페이지"],
+ "LAR9ub": [["0"], "의 기존 OpenKnowledge 항목을 대체합니다"],
"LBv_TR": ["위치를 심볼릭 링크로 변환"],
"LBwUR6": [
"브랜치 전환 시간이 초과되었습니다. 직접 ",
@@ -1705,6 +1707,9 @@
"OETysD": [
"아직 허브 페이지가 없습니다. 페이지에 들어오는 그래프 링크가 쌓이면 허브가 표시됩니다."
],
+ "OFIL46": [
+ "사용 중인 AI 도구에 프로젝트를 읽고 업데이트할 권한을 부여하세요. 이 설정은 나중에 Settings > AI tools & CLI에서 변경할 수 있습니다."
+ ],
"OJ4VDl": [
"이 에이전트는 ",
["0"],
@@ -2032,7 +2037,6 @@
"TDeOpa": [
"협업자와 에이전트의 편집은 유지하면서 직접 편집한 내용을 실행 취소하거나 다시 실행합니다."
],
- "TEEPPF": ["기존 OpenKnowledge 항목 교체 예정"],
"TEJiKw": ["OpenKnowledge에서 ", ["subject"], " 확인"],
"TKQ7K-": ["설치"],
"TLIO3j": ["기울임꼴 서식을 전환합니다."],
@@ -2622,6 +2626,7 @@
"이 함께 들어 있습니다. 플러그인에서 원하는 항목을 설치하세요."
],
"aRI3Om": ["창 닫기"],
+ "aSVmqV": ["아직 감지된 AI 도구가 없습니다. 설치한 후 Settings > This project에서 연결하세요."],
"aWlJai": ["로그인 흐름 시작 중"],
"aXkLuE": ["잘못된 apply 응답"],
"a_2_uB": [["trimmed"], " 생성 실패: ", ["0"]],
@@ -2820,7 +2825,6 @@
"dVaXnv": ["콘텐츠 규칙을 아직 불러오지 못했습니다. 잠시 후 다시 시도합니다."],
"dXCi3J": ["문제가 발생했습니다"],
"dXOolp": ["변경 사항 동기화 중"],
- "dXWMFn": ["편집기에서 스킬 제거"],
"darW93": [
[
"pageCount",
@@ -2877,6 +2881,7 @@
"ePK91l": ["편집"],
"eQkgKV": ["설치됨"],
"eQn7nI": ["설정 공유 로드 중"],
+ "eQsWFc": ["에이전트 스킬"],
"eQxm13": ["첨부 파일 위치를 업데이트하지 못했습니다: ", ["detail"]],
"eRSPr9": ["터미널 열기"],
"eTL5y9": ["선택 영역 미리보기 보기"],
@@ -2960,6 +2965,7 @@
"fhaxJL": ["클론 실패, 응답 본문이 비어 있습니다"],
"fiMNLl": ["동기화 설정이 아직 로드되지 않았습니다. 잠시 후 다시 시도하세요."],
"fj5YiP": ["GitHub에서 최신 사본을 다운로드합니다."],
+ "fmwwju": ["각 도구에 OpenKnowledge의 MCP 연결을 추가합니다."],
"fnlBN4": ["자동 승인 설정을 아직 불러오지 못했습니다. 잠시 후 다시 시도합니다."],
"fo0VXg": ["설치 제거"],
"fo6CCk": ["패널 ", ["0"]],
@@ -3121,7 +3127,6 @@
"hmYd9H": [
"함께 일하는 사람과 회사, 그들과의 회의를 기록합니다. 누군지, 어떻게 알게 되었는지, 마지막으로 무엇을 이야기했는지 기억하는 데 유용합니다."
],
- "hnlGzG": ["지금 건너뛰기"],
"hnq8fP": ["컨텍스트 항목 ", ["label"], " 제거"],
"hp8OtS": ["추가됨"],
"hpDtUm": ["외부 위키 링크"],
@@ -3235,6 +3240,7 @@
"j2kt1j": [
"선호하는 AI로 새 세션을 엽니다. 텍스트를 선택한 경우 실행하는 대신 해당 세션에 추가합니다."
],
+ "j4SmVY": [["0"], "을(를) 이 프로젝트에 연결"],
"j5CWO4": ["최신 상태"],
"j9Vqvc": ["상속됨"],
"j9tqwO": ["사이드바에 없음"],
@@ -3263,6 +3269,7 @@
"jTLWxh": [
"외부 페이지를 인라인 iframe에 삽입합니다. 문서, 데모, Figma, CodeSandbox를 지원합니다."
],
+ "jWzDfC": ["보유한 AI 도구를 확인하는 중"],
"jZ_E5M": ["가져오기 실패. 자세한 내용은 서버 로그를 확인하세요."],
"j_Z6g7": ["공유 기본값"],
"jbSG8V": ["파일 보기 (", ["sidebarShortcutLabel"], ")"],
@@ -3387,9 +3394,6 @@
"kpqWwG": [["targetNoun"], " 삭제됨: <0>", ["branch"], "</0> 브랜치"],
"ktspdq": ["파일을 열지 않고 lint 또는 링크 문제가 있는 파일에 색조와 배지를 적용합니다."],
"ku0yLQ": ["검색에 필요할 때 페이지를 처음 임베딩합니다."],
- "kwLcuN": [
- "사용할 AI 도구에 프로젝트 읽기 및 업데이트 권한 부여. 아래에서 설정하고 설정 > AI 도구 및 CLI에서 언제든 변경할 수 있습니다."
- ],
"kwNS82": ["파일을 수정할 때까지 이 테마를 사용할 수 없습니다."],
"ky9Tuc": [["0"], " 소스 편집"],
"kzpkcn": ["편집 내용 저장에 실패했습니다."],
@@ -3976,6 +3980,9 @@
"tHKQ6j": ["이름 변경 중"],
"tKlWWY": ["이모지"],
"tLX6LY": ["OpenKnowledge에서 문서 작성"],
+ "tLd2qa": [
+ "아직 감지된 AI 도구가 없습니다. 설치한 후 Settings > AI tools & CLI에서 연결하세요."
+ ],
"tM5uzl": [
"출처, 신뢰, 수명 주기 그룹은 sources, generated, verified, status, stale_after입니다. 모든 필드는 선택 사항이며, 없어도 오류가 아닙니다."
],
@@ -4239,6 +4246,7 @@
"xEb24l": ["이메일로 후속 연락 받기"],
"xFQRu4": ["이미 로컬에 있음"],
"xG4tb6": ["\"", ["docName"], "\" 연결이 끊겼습니다."],
+ "xGVfLh": ["계속"],
"xHGM4U": ["HTML 미리보기 숨기기"],
"xK291b": ["두 페이지 (홀수)"],
"xKDsGS": ["줄 ", ["0"], ", ", ["1"], "에서 이동"],
@@ -4323,6 +4331,9 @@
"ycIX5i": ["로그아웃된 상태입니다. 동기화를 계속하려면 로그인하세요."],
"yckias": ["파일 이름 변경"],
"ydzS9x": ["종료"],
+ "yeMZk8": [
+ "각 도구에 OpenKnowledge의 MCP 연결을 추가하고, 에이전트가 OpenKnowledge 프로젝트를 인식하도록 open-knowledge-discovery 스킬도 함께 추가합니다."
+ ],
"yf5auF": ["ID, 별칭 또는 이름으로 규칙 검색"],
"yf5r22": [
["keyName"],
diff --git a/packages/app/src/locales/ko/messages.po b/packages/app/src/locales/ko/messages.po
index f79fe192b..8f1e94150 100644
--- a/packages/app/src/locales/ko/messages.po
+++ b/packages/app/src/locales/ko/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, other {#개 문서 전체에 문제가 없습니다.}}"
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, other {플러그인}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, other {도구 #개 더 보기}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "관리 블록을 다음 대상에 추가합니다"
msgid "Adds a managed block to {0}"
msgstr "{0}에 관리 블록 추가"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "OpenKnowledge MCP 항목과 프로젝트 스킬을 추가하여 각 도구가 이 문서를 읽고 업데이트할 수 있게 합니다."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "각 도구에 OpenKnowledge의 MCP 연결을 추가하고, 에이전트가 OpenKnowledge 프로젝트를 인식하도록 open-knowledge-discovery 스킬도 함께 추가합니다."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "각 도구에 OpenKnowledge의 MCP 연결을 추가합니다."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "사용 가능한 경우 텔레메트리, 서버 상태, 런타임 정보를 추가합니다. 자격 증명은 항상 제거되며, 포함된 문서 이름은 비공개 처리되지 않은 일반 텍스트로 표시됩니다."
@@ -1469,9 +1479,12 @@ msgstr "Terminal을 읽기 쉽게 에이전트 패널을 닫았습니다."
msgid "Agent settings"
msgstr "에이전트 설정"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "에이전트 스킬"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "에이전트 스킬"
@@ -1488,6 +1501,8 @@ msgstr "에이전트"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI 도구 관리는 OpenKnowledge 데스크톱 앱에서만 사용할 수 있습니다."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "스킬 확인 중"
msgid "Checking source"
msgstr "소스 확인 중"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "보유한 AI 도구를 확인하는 중"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "GitHub 연결 확인 중"
@@ -2925,12 +2944,15 @@ msgstr "충돌하는 파일: {0}."
msgid "Conflicts"
msgstr "충돌"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "연결"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "{0}을(를) 이 프로젝트에 연결"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "GitHub 계정을 연결하면 저장소를 찾아보고 동기화할 수 있습니다."
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "공유하려면 이 프로젝트를 GitHub에 연결합니다."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "AI 도구 연결"
@@ -3063,7 +3084,6 @@ msgstr "콘텐츠 규칙"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "콘텐츠 규칙을 아직 불러오지 못했습니다. 잠시 후 다시 시도합니다."
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "컨텍스트 비용"
@@ -3076,6 +3096,11 @@ msgstr "컨텍스트 창:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "컨텍스트 창: {percent}% 사용됨, {usedLabel} / {sizeLabel} 토큰 사용"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "계속"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "브라우저에서 계속"
@@ -4661,10 +4686,6 @@ msgstr "예: 큰 표를 붙여넣은 후 편집기가 멈춤"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "각 팩은 바로 사용할 수 있는 폴더와 템플릿으로 프로젝트의 뼈대를 만듭니다."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "선택한 각 도구에 OpenKnowledge MCP 항목이 추가됩니다."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "유효한 URL을 입력하세요. 예: https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "터널이 제공하는 공개 오리진을 입력하세요. 예: https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "필드"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "사용하는 AI 도구에 프로젝트를 읽고 업데이트할 권한을 부여합니다. 확인란을 선택하면 즉시 설정하고, 선택을 해제하면 제거합니다."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "사용할 AI 도구에 프로젝트 읽기 및 업데이트 권한 부여. 아래에서 설정하고 설정 > AI 도구 및 CLI에서 언제든 변경할 수 있습니다."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "사용 중인 AI 도구에 프로젝트를 읽고 업데이트할 권한을 부여하세요. 이 설정은 나중에 Settings > AI tools & CLI에서 변경할 수 있습니다."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "고급 설정 숨기기"
msgid "Hide Agents"
msgstr "에이전트 가리기"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "세부 정보 숨기기"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "수동으로 푸시하는 방법"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "설정 방법"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "{0} 설정 방법(브라우저에서 열림)"
@@ -6795,7 +6815,6 @@ msgstr "설치 중"
msgid "Installing {agentName}…"
msgstr "{agentName} 설치 중…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "설치 위치"
@@ -7438,7 +7457,7 @@ msgstr "MCP 자동 복구 실패"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "선택한 에이전트가 없습니다. 스킬은 저장되지만 어디
msgid "No AI agents installed"
msgstr "설치된 AI 에이전트 없음"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "아직 감지된 AI 도구가 없습니다. 설치한 후 Settings > AI tools & CLI에서 연결하세요."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "아직 감지된 AI 도구가 없습니다. 설치한 후 Settings > This project에서 연결하세요."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "AI 도구가 감지되지 않았습니다. 이 스킬을 사용하려면 하나를 설치하세요."
@@ -10179,11 +10207,6 @@ msgstr "최근 프로젝트에서 \"{projectName}\"을 제거했습니다. 폴
msgid "Removes from"
msgstr "제거 위치"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "편집기에서 스킬 제거"
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "제거 중"
@@ -10338,6 +10361,11 @@ msgstr "이 복사본을 소스를 가리키는 심볼릭 링크로 바꿀까요
msgid "Replace these symlinks with independent copies?"
msgstr "이 심볼릭 링크를 독립적인 복사본으로 바꿀까요?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "{0}의 기존 OpenKnowledge 항목을 대체합니다"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "플러그인 아래의 자체 페이지에서 설정합니다."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "이 프로젝트를 처음 여는 사용자의 동기화 기본값을 설정합니다. 이 설정은 저장소에 커밋됩니다."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "{0} 설정"
@@ -11405,9 +11433,7 @@ msgstr "문서 패널 표시"
msgid "Show external URL nodes"
msgstr "외부 URL 노드 보기"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "더 적게 보기"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "이제 스킬을 공유합니다. 팀과 공유하려면 .ok/skills를 커밋하세요."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "스킬은 이 컴퓨터에서 감지된 모든 AI 도구에 설치되며, 위에서 선택한 MCP 연결과는 무관합니다."
@@ -11655,11 +11679,6 @@ msgstr "skills.sh를 사용할 수 없어 GitHub 결과를 보기로 전환합
msgid "Skip"
msgstr "건너뛰기"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "지금 건너뛰기"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "AI 작업 내용? (선택 사항)"
msgid "What the agent may read, write, and surface."
msgstr "에이전트가 읽고, 쓰고, 표시할 수 있는 항목입니다."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "변경 내용"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "다음 위치에 생성됩니다: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "<0>{sanitized}</0>으로 저장됩니다."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "기존 OpenKnowledge 항목 교체 예정"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/pseudo/messages.json b/packages/app/src/locales/pseudo/messages.json
index 632b596f4..46d79a2fa 100644
--- a/packages/app/src/locales/pseudo/messages.json
+++ b/packages/app/src/locales/pseudo/messages.json
@@ -308,13 +308,6 @@
"3pU8e-": ["Ƥĩń ţàƀ"],
"3qDsdf": ["Ĩńśţàĺĺ śķĩĺĺ"],
"3ri_4O": ["Ķēēƥś ŷōũŕ ķńōŵĺēďĝē ƀàśē àĺĩĝńēď ŵĩţĥ ţĥē Ōƥēń Ķńōŵĺēďĝē Ƒōŕḿàţ."],
- "3s-k7L": [
- [
- "0",
- "plural",
- { "one": ["Śĥōŵ ", "#", " ḿōŕē ţōōĺ"], "other": ["Śĥōŵ ", "#", " ḿōŕē ţōōĺś"] }
- ]
- ],
"3uSzeS": ["Ĩḿàĝēś ƒŕōḿ àńŷŵĥēŕē · ńōń-ĩḿàĝē ƒĩĺēś ḿũśţ ƀē ĩń ţĥē ŵōŕķśƥàćē"],
"3v62tp": ["Ƥōƥũĺàŕ ōń śķĩĺĺś.śĥ"],
"3wiBAs": ["Ŵĥēŕē śĥōũĺď ĩţ ĺĩvē?"],
@@ -626,7 +619,6 @@
["0"],
" — ţĥàţ ƒōĺďēŕ ĩś ńōţ ţĥĩś śķĩĺĺ. Ŕēńàḿē ōŕ ďēĺēţē ōńē ōƒ ţĥē ţŵō ţō ŕēśōĺvē."
],
- "8IP5jY": ["Ēàćĥ śēĺēćţēď ţōōĺ ĝēţś àń ŌƥēńĶńōŵĺēďĝē ḾĆƤ ēńţŕŷ."],
"8Jm2EP": ["Ţōĝĝĺē ĥēàďĩńĝ ĺēvēĺś 1 ţĥŕōũĝĥ 6."],
"8L0EPw": ["ōŕ ƥàśţē à ţōķēń"],
"8OGgNM": ["Ōƥēń ōŕ ćĺōśē vĩśũàĺ-ēďĩţōŕ ƒĩńď."],
@@ -881,6 +873,9 @@
"C0cll5": ["Ĩńśţàĺĺēď àţ ", ["0"], " — ćĺĩćķ ţō ŕēḿōvē"],
"C0hoX6": ["Àĺŵàŷś ĩńćĺũďēď"],
"C3Zc44": ["Ēxƥĺōŕē śķĩĺĺś"],
+ "C4HK3A": [
+ "Àďďś àń ŌƥēńĶńōŵĺēďĝē ḾĆƤ ēńţŕŷ àńď ţĥē ƥŕōĴēćţ śķĩĺĺ, śō ēàćĥ ţōōĺ ćàń ŕēàď àńď ũƥďàţē ţĥēśē ńōţēś."
+ ],
"C62P6b": ["Ēďĩţōŕ śēĺēćţĩōń"],
"C7PKTv": ["Śēńď ćĥēćķēď ćōḿḿēńţś ţō ćĥàţ"],
"C7waqI": [["0", "plural", { "one": ["#", " ďōćũḿēńţ"], "other": ["#", " ďōćũḿēńţś"] }]],
@@ -1515,6 +1510,7 @@
"L83DKQ": ["ƤÀŢĤ ĩńśţàĺĺ ƒàĩĺēď: ", ["summary"]],
"L8fKuc": ["Ŕũń ţĥĩś ĩń ŷōũŕ ţēŕḿĩńàĺ"],
"L9Y9Om": ["Ţōƥ ĺĩńķēď ƥàĝēś"],
+ "LAR9ub": ["Ŕēƥĺàćēś ţĥē ēxĩśţĩńĝ ŌƥēńĶńōŵĺēďĝē ēńţŕŷ ĩń ", ["0"]],
"LBv_TR": ["Ćōńvēŕţ ĺōćàţĩōńś ţō śŷḿĺĩńķś"],
"LBwUR6": ["ßŕàńćĥ śŵĩţćĥ ţĩḿēď ōũţ — ţŕŷ ōƥēńĩńĝ ţĥē ", ["targetNoun"], " ḿàńũàĺĺŷ."],
"LCUy_N": ["Śēśśĩōń ēńďēď"],
@@ -1766,6 +1762,9 @@
"ODGGlj": ["Ƒōĺďēŕ ńàḿē: ", ["detail"]],
"OEEQKT": ["Ĝōōď"],
"OETysD": ["Ńō ĥũƀ ƥàĝēś ŷēţ. Ĥũƀś àƥƥēàŕ ōńćē ƥàĝēś àććũḿũĺàţē ĩńƀōũńď ĝŕàƥĥ ĺĩńķś."],
+ "OFIL46": [
+ "Ĝĩvē ţĥē ÀĨ ţōōĺś ŷōũ ũśē àććēśś ţō ŕēàď àńď ũƥďàţē ŷōũŕ ƥŕōĴēćţś. Ŷōũ ćàń ćĥàńĝē àńŷ ōƒ ţĥĩś ĺàţēŕ ĩń Śēţţĩńĝś > ÀĨ ţōōĺś & ĆĹĨ."
+ ],
"OJ4VDl": [
"Ţĥĩś àĝēńţ ŕũńś ţĥŕōũĝĥ ",
["0"],
@@ -2113,7 +2112,6 @@
"TDeOpa": [
"Ũńďō ōŕ ŕēďō ŷōũŕ ōŵń ēďĩţś, ĺēàvĩńĝ ēďĩţś ƒŕōḿ ćōĺĺàƀōŕàţōŕś àńď àĝēńţś ĩń ƥĺàćē."
],
- "TEEPPF": ["Ŵĩĺĺ ŕēƥĺàćē ēxĩśţĩńĝ ŌƥēńĶńōŵĺēďĝē ēńţŕŷ"],
"TEJiKw": ["ŌƥēńĶńōŵĺēďĝē ćĥēćķēď ", ["subject"]],
"TKQ7K-": ["Ĩńśţàĺĺ"],
"TLIO3j": ["Ţōĝĝĺē ĩţàĺĩć ƒōŕḿàţţĩńĝ."],
@@ -2716,6 +2714,9 @@
". Ĩńśţàĺĺ àńŷ ōƒ ţĥēḿ ƒŕōḿ ţĥē ƥĺũĝĩń."
],
"aRI3Om": ["Ćĺōśē Ŵĩńďōŵ"],
+ "aSVmqV": [
+ "Ńō ÀĨ ţōōĺś ďēţēćţēď ŷēţ. Ōńćē ŷōũ ĩńśţàĺĺ ōńē, ćōńńēćţ ĩţ ƒŕōḿ Śēţţĩńĝś > Ţĥĩś ƥŕōĴēćţ."
+ ],
"aWlJai": ["Śţàŕţĩńĝ śĩĝń-ĩń ƒĺōŵ"],
"aXkLuE": ["Ḿàĺƒōŕḿēď àƥƥĺŷ ŕēśƥōńśē"],
"a_2_uB": ["Ćōũĺďń'ţ ćŕēàţē ", ["trimmed"], ": ", ["0"]],
@@ -2919,7 +2920,6 @@
"dVaXnv": ["Ćōńţēńţ ŕũĺēś ńōţ ŷēţ ĺōàďēď — ţŕŷ àĝàĩń ĩń à ḿōḿēńţ"],
"dXCi3J": ["Śōḿēţĥĩńĝ ƀŕōķē"],
"dXOolp": ["Śŷńćĩńĝ ŷōũŕ ćĥàńĝēś"],
- "dXWMFn": ["Ŕēḿōvēś ţĥĩś śķĩĺĺ ƒŕōḿ ŷōũŕ ēďĩţōŕś."],
"darW93": [
[
"pageCount",
@@ -2982,6 +2982,7 @@
"ePK91l": ["Ēďĩţ"],
"eQkgKV": ["Ĩńśţàĺĺēď"],
"eQn7nI": ["Ĺōàďĩńĝ ćōńƒĩĝ śĥàŕĩńĝ"],
+ "eQsWFc": ["Àĝēńţ Śķĩĺĺ"],
"eQxm13": ["Ƒàĩĺēď ţō ũƥďàţē àţţàćĥḿēńţ ĺōćàţĩōń — ", ["detail"]],
"eRSPr9": ["Ōƥēń ţēŕḿĩńàĺ"],
"eTL5y9": ["Śĥōŵ śēĺēćţĩōń ƥŕēvĩēŵ"],
@@ -3061,6 +3062,7 @@
"fhaxJL": ["Ćĺōńē ƒàĩĺēď — ēḿƥţŷ ŕēśƥōńśē ƀōďŷ"],
"fiMNLl": ["Śŷńć śēţţĩńĝś ńōţ ŷēţ ĺōàďēď — ţŕŷ àĝàĩń ĩń à ḿōḿēńţ"],
"fj5YiP": ["Ďōŵńĺōàďś à ƒŕēśĥ ćōƥŷ ƒŕōḿ ĜĩţĤũƀ."],
+ "fmwwju": ["Àďďś ŌƥēńĶńōŵĺēďĝē'ś ḾĆƤ ćōńńēćţĩōń ţō ēàćĥ."],
"fnlBN4": ["Àũţō-àƥƥŕōvē śēţţĩńĝś ńōţ ŷēţ ĺōàďēď — ţŕŷ àĝàĩń ĩń à ḿōḿēńţ"],
"fo0VXg": ["Ũńĩńśţàĺĺ"],
"fo6CCk": ["ƥàńē ", ["0"]],
@@ -3235,7 +3237,6 @@
"hmYd9H": [
"Ķēēƥ ţŕàćķ ōƒ ţĥē ƥēōƥĺē àńď ćōḿƥàńĩēś ŷōũ ŵōŕķ ŵĩţĥ àńď ţĥē ḿēēţĩńĝś ŷōũ ĥàvē ŵĩţĥ ţĥēḿ. Ĝōōď ƒōŕ ŕēḿēḿƀēŕĩńĝ ŵĥō śōḿēōńē ĩś, ĥōŵ ŷōũ ķńōŵ ţĥēḿ, àńď ŵĥàţ ŷōũ ĺàśţ ţàĺķēď àƀōũţ."
],
- "hnlGzG": ["Śķĩƥ ƒōŕ ńōŵ"],
"hnq8fP": ["Ŕēḿōvē ", ["label"], " ƒŕōḿ ćōńţēxţ"],
"hp8OtS": ["Àďďēď"],
"hpDtUm": ["Ēxţēŕńàĺ ŵĩķĩ ĺĩńķ"],
@@ -3352,6 +3353,7 @@
"j2kt1j": [
"Ōƥēń à ńēŵ śēśśĩōń ŵĩţĥ ŷōũŕ ƥŕēƒēŕŕēď ÀĨ. Ŵĩţĥ ţēxţ śēĺēćţēď, śţàĝē ĩţ ţĥēŕē ĩńśţēàď ōƒ ŕũńńĩńĝ ĩţ."
],
+ "j4SmVY": ["Ćōńńēćţ ", ["0"], " ţō ţĥĩś ƥŕōĴēćţ"],
"j5CWO4": ["Ũƥ ţō ďàţē"],
"j9Vqvc": ["ĩńĥēŕĩţēď"],
"j9tqwO": ["Ńōţ ĩń śĩďēƀàŕ"],
@@ -3376,6 +3378,7 @@
"jSO1X5": ["Ĩńĥēŕĩţēď ţēḿƥĺàţēś ćàń'ţ ƀē ŕēńàḿēď ĥēŕē — ŕēńàḿē ĩţ ĩń ţĥē ƒōĺďēŕ ţĥàţ ōŵńś ĩţ."],
"jSY41h": ["\"", ["docName"], "\" ţōōķ ţōō ĺōńĝ. Ćĥēćķ ŷōũŕ ćōńńēćţĩōń."],
"jTLWxh": ["Ēḿƀēď àń ēxţēŕńàĺ ƥàĝē ĩń àń ĩńĺĩńē ĩƒŕàḿē (ďōćś, ďēḿōś, Ƒĩĝḿà, ĆōďēŚàńďƀōx)."],
+ "jWzDfC": ["Ćĥēćķĩńĝ ŵĥĩćĥ ÀĨ ţōōĺś ŷōũ ĥàvē"],
"jZ_E5M": ["Ƒēţćĥ ƒàĩĺēď — ćĥēćķ ţĥē śēŕvēŕ ĺōĝś ƒōŕ ďēţàĩĺś."],
"j_Z6g7": ["Śĥàŕēď ďēƒàũĺţ"],
"jbSG8V": ["Śĥōŵ Ƒĩĺēś (", ["sidebarShortcutLabel"], ")"],
@@ -3507,9 +3510,6 @@
"kpqWwG": ["Ţĥĩś ", ["targetNoun"], " ŵàś ŕēḿōvēď ƒŕōḿ ƀŕàńćĥ <0>", ["branch"], "</0>."],
"ktspdq": ["Ţĩńţ àńď ƀàďĝē ƒĩĺēś ţĥàţ ĥàvē ĺĩńţ ōŕ ĺĩńķ ƥŕōƀĺēḿś, ŵĩţĥōũţ ōƥēńĩńĝ ţĥēḿ."],
"ku0yLQ": ["Ƥàĝēś àŕē ēḿƀēďďēď ţĥē ƒĩŕśţ ţĩḿē à śēàŕćĥ ńēēďś ţĥēḿ."],
- "kwLcuN": [
- "Ĝĩvē ţĥē ÀĨ ţōōĺś ŷōũ ũśē àććēśś ţō ŕēàď àńď ũƥďàţē ŷōũŕ ƥŕōĴēćţś. Ƥĩćķ ŵĥàţ ţō śēţ ũƥ ƀēĺōŵ, àńď ćĥàńĝē ĩţ àńŷţĩḿē ĩń Śēţţĩńĝś > ÀĨ ţōōĺś & ĆĹĨ."
- ],
"kwNS82": ["Ţĥĩś ţĥēḿē ćàń’ţ ƀē ũśēď ũńţĩĺ ţĥē ƒĩĺē ĩś ƒĩxēď."],
"ky9Tuc": ["Ēďĩţ ", ["0"], " śōũŕćē"],
"kzpkcn": ["Ćōũĺďń'ţ śàvē ţĥàţ ēďĩţ."],
@@ -4134,6 +4134,9 @@
"tHKQ6j": ["Ŕēńàḿĩńĝ"],
"tKlWWY": ["ĒḿōĴĩ"],
"tLX6LY": ["ŌƥēńĶńōŵĺēďĝē ŵŕōţē à ďōćũḿēńţ"],
+ "tLd2qa": [
+ "Ńō ÀĨ ţōōĺś ďēţēćţēď ŷēţ. Ōńćē ŷōũ ĩńśţàĺĺ ōńē, ćōńńēćţ ĩţ ƒŕōḿ Śēţţĩńĝś > ÀĨ ţōōĺś & ĆĹĨ."
+ ],
"tM5uzl": [
"Ţĥē ƥŕōvēńàńćē, ţŕũśţ, àńď ĺĩƒēćŷćĺē ƒàḿĩĺĩēś: śōũŕćēś, ĝēńēŕàţēď, vēŕĩƒĩēď, śţàţũś, àńď śţàĺē_àƒţēŕ. Ēvēŕŷ ƒĩēĺď ĩś ōƥţĩōńàĺ, àńď àƀśēńćē ĩś ńēvēŕ àń ēŕŕōŕ."
],
@@ -4428,6 +4431,7 @@
"xEb24l": ["Ĺēţ ũś ƒōĺĺōŵ ũƥ ƀŷ ēḿàĩĺ"],
"xFQRu4": ["Ĩ àĺŕēàďŷ ĥàvē ĩţ ĺōćàĺĺŷ"],
"xG4tb6": ["Ĺōśţ ćōńńēćţĩōń ţō \"", ["docName"], "\"."],
+ "xGVfLh": ["Ćōńţĩńũē"],
"xHGM4U": ["Ĥĩďē ĤŢḾĹ ƥŕēvĩēŵ"],
"xK291b": ["Ţŵō-ƥàĝē (ōďď)"],
"xKDsGS": ["Ĝō ţō ĺĩńē ", ["0"], " ĩń ", ["1"]],
@@ -4519,6 +4523,9 @@
"ycIX5i": ["Ŷōũ'ŕē śĩĝńēď ōũţ — śĩĝń ĩń ţō ŕēśũḿē śŷńćĩńĝ."],
"yckias": ["Ŕēńàḿē ƒĩĺē"],
"ydzS9x": ["Ēxĩţ"],
+ "yeMZk8": [
+ "Àďďś ŌƥēńĶńōŵĺēďĝē'ś ḾĆƤ ćōńńēćţĩōń ţō ēàćĥ, ƥĺũś ţĥē ōƥēń-ķńōŵĺēďĝē-ďĩśćōvēŕŷ śķĩĺĺ śō ŷōũŕ àĝēńţś ŕēćōĝńĩźē ŌƥēńĶńōŵĺēďĝē ƥŕōĴēćţś."
+ ],
"yf5auF": ["Śēàŕćĥ ŕũĺēś ƀŷ ĩď, àĺĩàś, ōŕ ńàḿē"],
"yf5r22": [
["keyName"],
diff --git a/packages/app/src/locales/pseudo/messages.po b/packages/app/src/locales/pseudo/messages.po
index 651ae9dcd..0a0af81e3 100644
--- a/packages/app/src/locales/pseudo/messages.po
+++ b/packages/app/src/locales/pseudo/messages.po
@@ -278,11 +278,6 @@ msgstr ""
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr ""
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr ""
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1410,6 +1405,21 @@ msgstr ""
msgid "Adds a managed block to {0}"
msgstr ""
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr ""
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr ""
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr ""
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr ""
@@ -1465,9 +1475,12 @@ msgstr ""
msgid "Agent settings"
msgstr ""
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr ""
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr ""
@@ -1484,6 +1497,8 @@ msgstr ""
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr ""
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2371,6 +2386,10 @@ msgstr ""
msgid "Checking source"
msgstr ""
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr ""
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr ""
@@ -2921,12 +2940,15 @@ msgstr ""
msgid "Conflicts"
msgstr ""
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr ""
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr ""
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr ""
@@ -2960,7 +2982,6 @@ msgid "Connect this project to GitHub to share."
msgstr ""
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr ""
@@ -3059,7 +3080,6 @@ msgstr ""
msgid "Content rules not yet loaded — try again in a moment"
msgstr ""
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr ""
@@ -3072,6 +3092,11 @@ msgstr ""
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr ""
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr ""
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr ""
@@ -4657,10 +4682,6 @@ msgstr ""
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr ""
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr ""
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5017,7 +5038,6 @@ msgstr ""
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr ""
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5527,7 +5547,6 @@ msgstr ""
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5977,7 +5996,7 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr ""
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
msgstr ""
#: src/components/settings/LintingSection.tsx
@@ -6160,7 +6179,11 @@ msgstr ""
msgid "Hide Agents"
msgstr ""
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr ""
@@ -6302,14 +6325,11 @@ msgid "How to push manually"
msgstr ""
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr ""
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr ""
@@ -6791,7 +6811,6 @@ msgstr ""
msgid "Installing {agentName}…"
msgstr ""
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr ""
@@ -7434,7 +7453,7 @@ msgstr ""
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8042,6 +8061,15 @@ msgstr ""
msgid "No AI agents installed"
msgstr ""
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr ""
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr ""
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr ""
@@ -10175,11 +10203,6 @@ msgstr ""
msgid "Removes from"
msgstr ""
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr ""
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr ""
@@ -10334,6 +10357,11 @@ msgstr ""
msgid "Replace these symlinks with independent copies?"
msgstr ""
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr ""
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11209,7 +11237,7 @@ msgstr ""
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr ""
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr ""
@@ -11401,9 +11429,7 @@ msgstr ""
msgid "Show external URL nodes"
msgstr ""
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr ""
@@ -11622,8 +11648,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr ""
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr ""
@@ -11651,11 +11675,6 @@ msgstr ""
msgid "Skip"
msgstr ""
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr ""
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13989,11 +14008,13 @@ msgstr ""
msgid "What the agent may read, write, and surface."
msgstr ""
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr ""
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14087,11 +14108,6 @@ msgstr ""
msgid "Will be saved as <0>{sanitized}</0>."
msgstr ""
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr ""
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/pt-BR/messages.json b/packages/app/src/locales/pt-BR/messages.json
index 5ff43b50c..90fe02457 100644
--- a/packages/app/src/locales/pt-BR/messages.json
+++ b/packages/app/src/locales/pt-BR/messages.json
@@ -353,17 +353,6 @@
"3pU8e-": ["Fixar aba"],
"3qDsdf": ["Instalar habilidade"],
"3ri_4O": ["Mantém sua base de conhecimento alinhada ao Open Knowledge Format."],
- "3s-k7L": [
- [
- "0",
- "plural",
- {
- "one": ["Mostrar mais ", "#", " ferramenta"],
- "many": ["Mostrar mais ", "#", " ferramentas"],
- "other": ["Mostrar mais ", "#", " ferramentas"]
- }
- ]
- ],
"3uSzeS": [
"Imagens de qualquer lugar · arquivos que não são imagens devem estar no espaço de trabalho"
],
@@ -734,7 +723,6 @@
["0"],
" — essa pasta não é esta habilidade. Renomeie ou exclua uma das duas para resolver."
],
- "8IP5jY": ["Cada ferramenta selecionada recebe uma entrada MCP do OpenKnowledge."],
"8Jm2EP": ["Liga ou desliga os níveis de título de 1 a 6."],
"8L0EPw": ["ou cole um token"],
"8OGgNM": ["Abre ou fecha a busca do editor visual."],
@@ -1018,6 +1006,9 @@
"C0cll5": ["Instalado em ", ["0"], " — clique para remover"],
"C0hoX6": ["Sempre incluído"],
"C3Zc44": ["Explorar habilidades"],
+ "C4HK3A": [
+ "Adiciona uma entrada MCP do OpenKnowledge e a habilidade do projeto, para que cada ferramenta possa ler e atualizar estes documentos."
+ ],
"C62P6b": ["Seleção do editor"],
"C7PKTv": ["Enviar os comentários marcados para o chat"],
"C7waqI": [
@@ -1748,6 +1739,7 @@
"L83DKQ": ["A instalação no PATH falhou: ", ["summary"]],
"L8fKuc": ["Rode isto no seu terminal"],
"L9Y9Om": ["Páginas mais ligadas"],
+ "LAR9ub": ["Substitui a entrada existente do OpenKnowledge em ", ["0"]],
"LBv_TR": ["Converter os locais em links simbólicos"],
"LBwUR6": ["A troca de ramo expirou — tente abrir ", ["targetNoun"], " manualmente."],
"LCUy_N": ["A sessão terminou"],
@@ -2022,6 +2014,9 @@
"OETysD": [
"Ainda não há páginas de índice. Os índices aparecem quando as páginas acumulam links de entrada no grafo."
],
+ "OFIL46": [
+ "Dê às ferramentas de IA que você usa acesso para ler e atualizar seus projetos. Você pode mudar tudo isso depois em Configurações > Ferramentas de IA e CLI."
+ ],
"OJ4VDl": [
"Este agente roda pelo ",
["0"],
@@ -2396,7 +2391,6 @@
"TDeOpa": [
"Desfaz ou refaz suas próprias edições, deixando intactas as de colaboradores e agentes."
],
- "TEEPPF": ["Vai substituir a entrada existente do OpenKnowledge"],
"TEJiKw": ["OpenKnowledge verificou ", ["subject"]],
"TKQ7K-": ["Instalar"],
"TLIO3j": ["Liga ou desliga o itálico."],
@@ -3083,6 +3077,9 @@
". Instale qualquer uma delas pelo plugin."
],
"aRI3Om": ["Fechar janela"],
+ "aSVmqV": [
+ "Nenhuma ferramenta de IA detectada ainda. Depois de instalar uma, conecte-a em Configurações > Este projeto."
+ ],
"aWlJai": ["Iniciando o fluxo de login"],
"aXkLuE": ["Resposta de aplicação malformada"],
"a_2_uB": ["Não deu para criar ", ["trimmed"], ": ", ["0"]],
@@ -3322,7 +3319,6 @@
"dVaXnv": ["As regras de conteúdo ainda não carregaram — tente de novo em instantes"],
"dXCi3J": ["Algo quebrou"],
"dXOolp": ["Sincronizando suas alterações"],
- "dXWMFn": ["Remove esta habilidade dos seus editores."],
"darW93": [
[
"pageCount",
@@ -3398,6 +3394,7 @@
"ePK91l": ["Editar"],
"eQkgKV": ["Instalado"],
"eQn7nI": ["Carregando compartilhamento da configuração"],
+ "eQsWFc": ["Habilidade de agente"],
"eQxm13": ["Falha ao atualizar o local dos anexos — ", ["detail"]],
"eRSPr9": ["Abrir o terminal"],
"eTL5y9": ["Mostrar a pré-visualização da seleção"],
@@ -3499,6 +3496,7 @@
"As configurações de sincronização ainda não carregaram — tente de novo em instantes"
],
"fj5YiP": ["Baixa uma cópia nova do GitHub."],
+ "fmwwju": ["Adiciona a conexão MCP do OpenKnowledge a cada uma."],
"fnlBN4": [
"As configurações de aprovação automática ainda não carregaram — tente de novo em instantes"
],
@@ -3686,7 +3684,6 @@
"hmYd9H": [
"Acompanhe as pessoas e empresas com quem você trabalha e as reuniões que tem com elas. Bom para lembrar quem é alguém, de onde você a conhece e sobre o que vocês falaram da última vez."
],
- "hnlGzG": ["Pular por enquanto"],
"hnq8fP": ["Remover ", ["label"], " do contexto"],
"hp8OtS": ["Adicionado"],
"hpDtUm": ["Link wiki externo"],
@@ -3819,6 +3816,7 @@
"j2kt1j": [
"Abre uma nova sessão com a IA que você preferir. Com um texto selecionado, prepara-o lá em vez de executar."
],
+ "j4SmVY": ["Conectar ", ["0"], " a este projeto"],
"j5CWO4": ["Atualizado"],
"j9Vqvc": ["herdado"],
"j9tqwO": ["Não está na barra lateral"],
@@ -3845,6 +3843,7 @@
"jTLWxh": [
"Incorpora uma página externa em um iframe embutido (documentação, demos, Figma, CodeSandbox)."
],
+ "jWzDfC": ["Verificando quais ferramentas de IA você tem"],
"jZ_E5M": ["Falha ao buscar — veja os logs do servidor para os detalhes."],
"j_Z6g7": ["Padrão compartilhado"],
"jbSG8V": ["Mostrar Arquivos (", ["sidebarShortcutLabel"], ")"],
@@ -3991,9 +3990,6 @@
"Colore e marca os arquivos com problemas de análise ou de links, sem precisar abri-los."
],
"ku0yLQ": ["As páginas viram embeddings na primeira vez que uma busca precisa delas."],
- "kwLcuN": [
- "Dê às ferramentas de IA que você usa acesso para ler e atualizar seus projetos. Escolha abaixo o que configurar e mude quando quiser em Configurações > Ferramentas de IA e CLI."
- ],
"kwNS82": ["Este tema não pode ser usado até que o arquivo seja corrigido."],
"ky9Tuc": ["Editar o código de ", ["0"]],
"kzpkcn": ["Não deu para salvar essa edição."],
@@ -4683,6 +4679,9 @@
"tHKQ6j": ["Renomeando"],
"tKlWWY": ["Emoji"],
"tLX6LY": ["OpenKnowledge escreveu um documento"],
+ "tLd2qa": [
+ "Nenhuma ferramenta de IA detectada ainda. Depois de instalar uma, conecte-a em Configurações > Ferramentas de IA e CLI."
+ ],
"tM5uzl": [
"As famílias de origem, confiança e ciclo de vida: sources, generated, verified, status e stale_after. Cada campo é opcional e a ausência nunca é um erro."
],
@@ -5002,6 +5001,7 @@
"xEb24l": ["Deixe a gente responder por e-mail"],
"xFQRu4": ["Já tenho isso localmente"],
"xG4tb6": ["A conexão com \"", ["docName"], "\" foi perdida."],
+ "xGVfLh": ["Continuar"],
"xHGM4U": ["Ocultar a pré-visualização do HTML"],
"xK291b": ["Duas páginas (ímpares)"],
"xKDsGS": ["Ir para a linha ", ["0"], " em ", ["1"]],
@@ -5096,6 +5096,9 @@
"ycIX5i": ["Você saiu da conta — entre para retomar a sincronização."],
"yckias": ["Renomear o arquivo"],
"ydzS9x": ["Sair"],
+ "yeMZk8": [
+ "Adiciona a conexão MCP do OpenKnowledge a cada uma, mais a habilidade open-knowledge-discovery para que seus agentes reconheçam projetos do OpenKnowledge."
+ ],
"yf5auF": ["Busque regras por id, apelido ou nome"],
"yf5r22": [
"Tipo de ",
diff --git a/packages/app/src/locales/pt-BR/messages.po b/packages/app/src/locales/pt-BR/messages.po
index 39cc3b1c2..aa26a20e4 100644
--- a/packages/app/src/locales/pt-BR/messages.po
+++ b/packages/app/src/locales/pt-BR/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {Nenhum problema em # documento.} many {Nenhum problema
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {Plugin} many {Plugins} other {Plugins}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {Mostrar mais # ferramenta} many {Mostrar mais # ferramentas} other {Mostrar mais # ferramentas}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1410,6 +1405,21 @@ msgstr "Adiciona um bloco gerenciado a"
msgid "Adds a managed block to {0}"
msgstr "Adiciona um bloco gerenciado a {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "Adiciona uma entrada MCP do OpenKnowledge e a habilidade do projeto, para que cada ferramenta possa ler e atualizar estes documentos."
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "Adiciona a conexão MCP do OpenKnowledge a cada uma, mais a habilidade open-knowledge-discovery para que seus agentes reconheçam projetos do OpenKnowledge."
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "Adiciona a conexão MCP do OpenKnowledge a cada uma."
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "Adiciona telemetria, estado do servidor e informações de execução quando disponíveis. As credenciais são sempre removidas; os nomes de documento, se incluídos, aparecem em texto simples (sem ocultação)."
@@ -1465,9 +1475,12 @@ msgstr "O painel de agentes foi fechado para manter o Terminal legível."
msgid "Agent settings"
msgstr "Configurações do agente"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "Habilidade de agente"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "Habilidades de agente"
@@ -1484,6 +1497,8 @@ msgstr "Agentes"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "O gerenciamento de ferramentas de IA só está disponível no app de desktop do OpenKnowledge."
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2371,6 +2386,10 @@ msgstr "Verificando habilidade"
msgid "Checking source"
msgstr "Verificando a fonte"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "Verificando quais ferramentas de IA você tem"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "Verificando sua conexão com o GitHub"
@@ -2921,12 +2940,15 @@ msgstr "Arquivos em conflito: {0}."
msgid "Conflicts"
msgstr "Conflitos"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "Conectar"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "Conectar {0} a este projeto"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "Conecte uma conta do GitHub para explorar e sincronizar seus repositórios."
@@ -2960,7 +2982,6 @@ msgid "Connect this project to GitHub to share."
msgstr "Conecte este projeto ao GitHub para compartilhar."
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "Conectar a ferramentas de IA"
@@ -3059,7 +3080,6 @@ msgstr "Regras de conteúdo"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "As regras de conteúdo ainda não carregaram — tente de novo em instantes"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "Custo de contexto"
@@ -3072,6 +3092,11 @@ msgstr "Janela de contexto:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "Janela de contexto: {percent}% usado, {usedLabel} de {sizeLabel} tokens"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "Continuar"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "Continuar no navegador"
@@ -4657,10 +4682,6 @@ msgstr "ex.: O editor travou depois que colei uma tabela grande"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "Cada pacote prepara seu projeto com pastas e modelos prontos."
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "Cada ferramenta selecionada recebe uma entrada MCP do OpenKnowledge."
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5017,7 +5038,6 @@ msgstr "Insira uma URL válida, por exemplo https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "Insira a origem pública que seu túnel serve, por exemplo https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5527,7 +5547,6 @@ msgstr "Campos"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5977,8 +5996,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "Dê às ferramentas de IA que você usa acesso para ler e atualizar seus projetos. Marcar uma caixa configura na hora; desmarcar remove."
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "Dê às ferramentas de IA que você usa acesso para ler e atualizar seus projetos. Escolha abaixo o que configurar e mude quando quiser em Configurações > Ferramentas de IA e CLI."
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "Dê às ferramentas de IA que você usa acesso para ler e atualizar seus projetos. Você pode mudar tudo isso depois em Configurações > Ferramentas de IA e CLI."
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6160,7 +6179,11 @@ msgstr "Ocultar as opções avançadas"
msgid "Hide Agents"
msgstr "Ocultar Agentes"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "Ocultar detalhes"
@@ -6302,14 +6325,11 @@ msgid "How to push manually"
msgstr "Como enviar manualmente"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "Como configurar"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "Como configurar {0} (abre no navegador)"
@@ -6791,7 +6811,6 @@ msgstr "Instalando"
msgid "Installing {agentName}…"
msgstr "Instalando {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "Instala em"
@@ -7434,7 +7453,7 @@ msgstr "O reparo automático do MCP falhou"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8042,6 +8061,15 @@ msgstr "Nenhum agente selecionado — as habilidades ficam salvas, mas não são
msgid "No AI agents installed"
msgstr "Nenhum agente de IA instalado"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "Nenhuma ferramenta de IA detectada ainda. Depois de instalar uma, conecte-a em Configurações > Ferramentas de IA e CLI."
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "Nenhuma ferramenta de IA detectada ainda. Depois de instalar uma, conecte-a em Configurações > Este projeto."
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "Nenhuma ferramenta de IA detectada. Instale uma para usar esta habilidade."
@@ -10175,11 +10203,6 @@ msgstr "\"{projectName}\" foi removido dos projetos recentes. A pasta dele não
msgid "Removes from"
msgstr "Remove de"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "Remove esta habilidade dos seus editores."
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "Removendo"
@@ -10334,6 +10357,11 @@ msgstr "Substituir estas cópias por links simbólicos para a fonte?"
msgid "Replace these symlinks with independent copies?"
msgstr "Substituir estes links simbólicos por cópias independentes?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "Substitui a entrada existente do OpenKnowledge em {0}"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11209,7 +11237,7 @@ msgstr "Configure na página própria dele, em Plugins."
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "Defina a sincronização padrão para quem abrir este projeto pela primeira vez. Esta configuração fica salva no seu repositório."
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "Configurar {0}"
@@ -11401,9 +11429,7 @@ msgstr "Mostrar o painel do documento"
msgid "Show external URL nodes"
msgstr "Mostrar os nós de URL externa"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "Mostrar menos"
@@ -11622,8 +11648,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "As habilidades agora são compartilhadas. Faça commit de .ok/skills para compartilhá-las com sua equipe."
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "As habilidades são instaladas em todas as ferramentas de IA detectadas nesta máquina, independentemente das conexões MCP que você escolheu acima."
@@ -11651,11 +11675,6 @@ msgstr "O skills.sh está indisponível; mostrando resultados do GitHub (sem cla
msgid "Skip"
msgstr "Pular"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "Pular por enquanto"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13989,11 +14008,13 @@ msgstr "O que a IA deve fazer? (opcional)"
msgid "What the agent may read, write, and surface."
msgstr "O que o agente pode ler, escrever e mostrar."
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "O que isto muda"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14087,11 +14108,6 @@ msgstr "Será criado em: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "Será salvo como <0>{sanitized}</0>."
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "Vai substituir a entrada existente do OpenKnowledge"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/ur/messages.json b/packages/app/src/locales/ur/messages.json
index 0787b0a6b..101c4ca03 100644
--- a/packages/app/src/locales/ur/messages.json
+++ b/packages/app/src/locales/ur/messages.json
@@ -307,9 +307,6 @@
"3pU8e-": ["ٹیب پن کریں"],
"3qDsdf": ["مہارت کو انسٹال کریں۔"],
"3ri_4O": ["آپ کے علمی ذخیرے کو Open Knowledge Format کے مطابق رکھتا ہے۔"],
- "3s-k7L": [
- ["0", "plural", { "one": ["#", " مزید ٹول دکھائیں"], "other": ["#", " مزید ٹولز دکھائیں"] }]
- ],
"3uSzeS": ["تصاویر کہیں سے بھی · غیر تصویری فائلیں ورک اسپیس میں ہونی چاہئیں"],
"3v62tp": ["skills.sh پر مقبول"],
"3wiBAs": ["یہ کہاں رہنا چاہیے؟"],
@@ -636,7 +633,6 @@
["0"],
" پر موجود ہے — وہ فولڈر یہ مہارت نہیں ہے۔ حل کے لیے دونوں میں سے کسی ایک کا نام بدلیں یا اسے حذف کریں۔"
],
- "8IP5jY": ["ہر منتخب ٹول کو ایک OpenKnowledge MCP اندراج ملتا ہے۔"],
"8Jm2EP": ["عنوان کی سطحیں 1 سے 6 تک بدلیں۔"],
"8L0EPw": ["یا کوئی ٹوکن چسپاں کریں"],
"8OGgNM": ["بصری ایڈیٹر کی تلاش کھولیں یا بند کریں۔"],
@@ -885,6 +881,9 @@
"C0cll5": [["0"], " پر نصب — ہٹانے کے لیے کلک کریں"],
"C0hoX6": ["ہمیشہ شامل"],
"C3Zc44": ["مہارتیں دریافت کریں"],
+ "C4HK3A": [
+ "OpenKnowledge کا MCP اندراج اور پروجیکٹ Skill شامل کرتا ہے، تاکہ ہر ٹول ان دستاویزات کو پڑھ اور اپ ڈیٹ کر سکے۔"
+ ],
"C62P6b": ["ایڈیٹر کا انتخاب"],
"C7PKTv": ["منتخب تبصرے چیٹ میں بھیجیں"],
"C7waqI": [["0", "plural", { "one": ["#", " دستاویز"], "other": ["#", " دستاویزات"] }]],
@@ -1547,6 +1546,7 @@
"L83DKQ": ["PATH انسٹال ناکام رہا: ", ["summary"]],
"L8fKuc": ["یہ اپنے ٹرمینل میں چلائیں"],
"L9Y9Om": ["سب سے زیادہ لنک شدہ صفحات"],
+ "LAR9ub": [["0"], " میں موجودہ OpenKnowledge اندراج کو بدل دے گا"],
"LBv_TR": ["مقامات کو سِم لنکس میں بدلیں"],
"LBwUR6": [
"برانچ تبدیل کرنے کا وقت ختم ہو گیا — ",
@@ -1806,6 +1806,9 @@
"OETysD": [
"ابھی کوئی مرکزی صفحہ نہیں۔ مراکز اسی وقت ظاہر ہوتے ہیں جب صفحات پر آنے والے گراف لنکس جمع ہونے لگیں۔"
],
+ "OFIL46": [
+ "آپ جو AI ٹولز استعمال کرتے ہیں انہیں اپنے پروجیکٹس پڑھنے اور اپ ڈیٹ کرنے کی رسائی دیں۔ یہ سب بعد میں ترتیبات > AI ٹولز اور CLI میں بدلا جا سکتا ہے۔"
+ ],
"OJ4VDl": [
"یہ ایجنٹ ",
["0"],
@@ -2167,7 +2170,6 @@
"TDeOpa": [
"اپنی ترامیم واپس لیں یا دوبارہ کریں، ساتھیوں اور ایجنٹس کی ترامیم اپنی جگہ چھوڑتے ہوئے۔"
],
- "TEEPPF": ["موجودہ OpenKnowledge اندراج کی جگہ لے لے گا"],
"TEJiKw": ["OpenKnowledge نے ", ["subject"], " کی جانچ کی"],
"TKQ7K-": ["نصب کریں"],
"TLIO3j": ["ترچھی فارمیٹنگ بدلیں۔"],
@@ -2779,6 +2781,9 @@
"۔ ان میں سے کوئی بھی پلگ اِن سے نصب کریں۔"
],
"aRI3Om": ["ونڈو بند کریں"],
+ "aSVmqV": [
+ "ابھی تک کوئی AI ٹول نہیں ملا۔ ایک انسٹال کرنے کے بعد اسے ترتیبات > یہ پروجیکٹ سے جوڑیں۔"
+ ],
"aWlJai": ["سائن اِن کا عمل شروع کیا جا رہا ہے"],
"aXkLuE": ["اپلائی کا جواب بگڑا ہوا ہے"],
"a_2_uB": [["trimmed"], " نہیں بن سکا: ", ["0"]],
@@ -2996,7 +3001,6 @@
"dVaXnv": ["مواد کے قواعد ابھی لوڈ نہیں ہوئے — تھوڑی دیر بعد دوبارہ کوشش کریں"],
"dXCi3J": ["کچھ ٹوٹ گیا"],
"dXOolp": ["آپ کی تبدیلیاں مطابق کی جا رہی ہیں"],
- "dXWMFn": ["یہ مہارت آپ کے ایڈیٹرز سے ہٹا دیتا ہے۔"],
"darW93": [
[
"pageCount",
@@ -3059,6 +3063,7 @@
"ePK91l": ["ترمیم"],
"eQkgKV": ["نصب شدہ"],
"eQn7nI": ["کنفیگ شیئرنگ لوڈ ہو رہی ہے"],
+ "eQsWFc": ["ایجنٹ مہارت"],
"eQxm13": ["منسلکات کا مقام تازہ نہیں ہو سکا — ", ["detail"]],
"eRSPr9": ["ٹرمینل کھولیں"],
"eTL5y9": ["انتخاب کا پیش منظر دکھائیں"],
@@ -3138,6 +3143,7 @@
"fhaxJL": ["کلون ناکام — جوابی مواد خالی ہے"],
"fiMNLl": ["مطابقت پذیری کی ترتیبات ابھی لوڈ نہیں ہوئیں — تھوڑی دیر بعد دوبارہ کوشش کریں"],
"fj5YiP": ["GitHub سے تازہ نقل ڈاؤن لوڈ کرتا ہے۔"],
+ "fmwwju": ["ہر ایک میں OpenKnowledge کا MCP کنکشن شامل کرتا ہے۔"],
"fnlBN4": ["خودکار منظوری کی ترتیبات ابھی لوڈ نہیں ہوئیں — تھوڑی دیر بعد دوبارہ کوشش کریں"],
"fo0VXg": ["ان انسٹال کریں"],
"fo6CCk": ["پین ", ["0"]],
@@ -3306,7 +3312,6 @@
"hmYd9H": [
"ان لوگوں اور کمپنیوں کا حساب رکھیں جن کے ساتھ آپ کام کرتے ہیں اور ان کے ساتھ ہونے والی ملاقاتوں کا۔ یہ یاد رکھنے کے لیے مفید کہ کوئی کون ہے، آپ اسے کیسے جانتے ہیں، اور آخری بار کس بارے میں بات ہوئی تھی۔"
],
- "hnlGzG": ["ابھی چھوڑ دیں"],
"hnq8fP": [["label"], " کو سیاق سے ہٹائیں"],
"hp8OtS": ["شامل کر دیا گیا"],
"hpDtUm": ["بیرونی وکی لنک"],
@@ -3428,6 +3433,7 @@
"j2kt1j": [
"اپنے پسندیدہ AI کے ساتھ نیا سیشن کھولیں۔ متن منتخب ہونے پر، اسے چلانے کے بجائے وہاں تیار رکھیں۔"
],
+ "j4SmVY": [["0"], " کو اس پروجیکٹ سے جوڑیں"],
"j5CWO4": ["تازہ ترین"],
"j9Vqvc": ["وراثت میں ملا"],
"j9tqwO": ["سائیڈ بار میں نہیں"],
@@ -3456,6 +3462,7 @@
"jTLWxh": [
"کوئی بیرونی صفحہ اِن لائن iframe میں شامل کریں (دستاویزات، ڈیمو، Figma، CodeSandbox)۔"
],
+ "jWzDfC": ["دیکھا جا رہا ہے کہ آپ کے پاس کون سے AI ٹولز ہیں"],
"jZ_E5M": ["لانا ناکام رہا — تفصیلات کے لیے سرور کے لاگز دیکھیں۔"],
"j_Z6g7": ["مشترکہ طے شدہ"],
"jbSG8V": ["فائلیں دکھائیں (", ["sidebarShortcutLabel"], ")"],
@@ -3589,9 +3596,6 @@
"kpqWwG": ["یہ ", ["targetNoun"], " برانچ <0>", ["branch"], "</0> سے ہٹا دیا گیا۔"],
"ktspdq": ["جن فائلوں میں لِنٹ یا لنک کے مسائل ہوں انہیں کھولے بغیر رنگ اور نشان دیں۔"],
"ku0yLQ": ["صفحات پہلی بار اسی وقت ایمبیڈ ہوتے ہیں جب کسی تلاش کو ان کی ضرورت پڑے۔"],
- "kwLcuN": [
- "اپنے استعمال کے AI ٹولز کو اپنے پروجیکٹس پڑھنے اور تازہ کرنے کی رسائی دیں۔ نیچے سے منتخب کریں کہ کیا ترتیب دینا ہے، اور اسے کسی بھی وقت ترتیبات > AI ٹولز اور CLI میں بدلیں۔"
- ],
"kwNS82": ["فائل درست کیے جانے تک یہ تھیم استعمال نہیں کی جا سکتی۔"],
"ky9Tuc": [["0"], " کا سورس ترمیم کریں"],
"kzpkcn": ["وہ ترمیم محفوظ نہیں ہو سکی۔"],
@@ -4210,6 +4214,9 @@
"tHKQ6j": ["نام بدلا جا رہا ہے"],
"tKlWWY": ["ایموجی"],
"tLX6LY": ["OpenKnowledge نے ایک دستاویز لکھی"],
+ "tLd2qa": [
+ "ابھی تک کوئی AI ٹول نہیں ملا۔ ایک انسٹال کرنے کے بعد اسے ترتیبات > AI ٹولز اور CLI سے جوڑیں۔"
+ ],
"tM5uzl": [
"اصل، اعتماد، اور لائف سائیکل فیملیز: sources، generated، verified، status، اور stale_after۔ ہر فیلڈ اختیاری ہے، اور غیر موجودگی کبھی بھی غلطی نہیں ہوتی۔"
],
@@ -4508,6 +4515,7 @@
"xEb24l": ["ہمیں ای میل کے ذریعے رابطہ کرنے دیں"],
"xFQRu4": ["یہ میرے پاس پہلے سے مقامی طور پر موجود ہے"],
"xG4tb6": ["«", ["docName"], "» سے رابطہ ختم ہو گیا۔"],
+ "xGVfLh": ["جاری رکھیں"],
"xHGM4U": ["HTML کا پیش منظر چھپائیں"],
"xK291b": ["دو صفحے (طاق)"],
"xKDsGS": [["1"], " میں سطر ", ["0"], " پر جائیں"],
@@ -4597,6 +4605,9 @@
"ycIX5i": ["آپ سائن آؤٹ ہیں — مطابقت پذیری جاری رکھنے کے لیے سائن اِن کریں۔"],
"yckias": ["فائل کا نام بدلیں"],
"ydzS9x": ["باہر نکلیں"],
+ "yeMZk8": [
+ "ہر ایک میں OpenKnowledge کا MCP کنکشن شامل کرتا ہے، ساتھ ہی open-knowledge-discovery Skill بھی، تاکہ آپ کے ایجنٹ OpenKnowledge پروجیکٹس پہچان سکیں۔"
+ ],
"yf5auF": ["قواعد کو شناخت، عرف یا نام سے تلاش کریں"],
"yf5r22": [
["keyName"],
diff --git a/packages/app/src/locales/ur/messages.po b/packages/app/src/locales/ur/messages.po
index 98ddbe0bd..1648fbefb 100644
--- a/packages/app/src/locales/ur/messages.po
+++ b/packages/app/src/locales/ur/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, one {# دستاویز میں کوئی مسئلہ نہیں۔}
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, one {پلگ اِن} other {پلگ اِنز}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, one {# مزید ٹول دکھائیں} other {# مزید ٹولز دکھائیں}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "منظم بلاک شامل کرتا ہے اِس میں"
msgid "Adds a managed block to {0}"
msgstr "{0} میں منظم بلاک شامل کرتا ہے"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "OpenKnowledge کا MCP اندراج اور پروجیکٹ Skill شامل کرتا ہے، تاکہ ہر ٹول ان دستاویزات کو پڑھ اور اپ ڈیٹ کر سکے۔"
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "ہر ایک میں OpenKnowledge کا MCP کنکشن شامل کرتا ہے، ساتھ ہی open-knowledge-discovery Skill بھی، تاکہ آپ کے ایجنٹ OpenKnowledge پروجیکٹس پہچان سکیں۔"
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "ہر ایک میں OpenKnowledge کا MCP کنکشن شامل کرتا ہے۔"
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "دستیاب ہونے پر ٹیلی میٹری، سرور کی حالت اور رن ٹائم معلومات شامل کرتا ہے۔ اسناد ہمیشہ ہٹا دی جاتی ہیں؛ دستاویزات کے نام، اگر شامل ہوں، سادہ متن میں ظاہر ہوتے ہیں (چھپائے نہیں جاتے)۔"
@@ -1469,9 +1479,12 @@ msgstr "Terminal کو پڑھنے کے قابل رکھنے کے لیے ایجنٹ
msgid "Agent settings"
msgstr "ایجنٹ کی ترتیبات"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "ایجنٹ مہارت"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "ایجنٹ مہارتیں"
@@ -1488,6 +1501,8 @@ msgstr "ایجنٹس"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI ٹولز کا انتظام صرف OpenKnowledge ڈیسک ٹاپ ایپ میں دستیاب ہے۔"
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "جانچ پڑتال کی مہارت"
msgid "Checking source"
msgstr "ماخذ جانچا جا رہا ہے"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "دیکھا جا رہا ہے کہ آپ کے پاس کون سے AI ٹولز ہیں"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "آپ کا GitHub کنکشن جانچا جا رہا ہے"
@@ -2925,12 +2944,15 @@ msgstr "متصادم فائلیں: {0}۔"
msgid "Conflicts"
msgstr "تنازعات"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "منسلک کریں"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "{0} کو اس پروجیکٹ سے جوڑیں"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "اپنی ریپازٹریز براؤز اور مطابق کرنے کے لیے GitHub اکاؤنٹ منسلک کریں۔"
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "شیئر کرنے کے لیے اس پروجیکٹ کو GitHub سے منسلک کریں۔"
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "AI ٹولز سے منسلک ہوں"
@@ -3063,7 +3084,6 @@ msgstr "مواد کے قواعد"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "مواد کے قواعد ابھی لوڈ نہیں ہوئے — تھوڑی دیر بعد دوبارہ کوشش کریں"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "کانٹیکسٹ لاگت"
@@ -3076,6 +3096,11 @@ msgstr "سیاق کی ونڈو:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "سیاق کی ونڈو: {percent}% استعمال شدہ، {sizeLabel} میں سے {usedLabel} ٹوکنز"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "جاری رکھیں"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "براؤزر میں جاری رکھیں"
@@ -4661,10 +4686,6 @@ msgstr "مثلاً بڑا ٹیبل چسپاں کرنے کے بعد ایڈیٹر
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "ہر پیک آپ کے پروجیکٹ کا ڈھانچہ تیار شدہ فولڈرز اور ٹیمپلیٹس سے بناتا ہے۔"
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "ہر منتخب ٹول کو ایک OpenKnowledge MCP اندراج ملتا ہے۔"
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "ایک درست URL درج کریں، مثلاً https://notes.example.c
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "وہ عوامی اوریجن درج کریں جو آپ کا ٹنل سرو کرتا ہے، مثلاً https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "فیلڈز"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "اپنے استعمال کے AI ٹولز کو اپنے پروجیکٹس پڑھنے اور تازہ کرنے کی رسائی دیں۔ خانہ نشان زد کرنا اسے فوراً ترتیب دے دیتا ہے؛ نشان ہٹانا اسے ہٹا دیتا ہے۔"
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "اپنے استعمال کے AI ٹولز کو اپنے پروجیکٹس پڑھنے اور تازہ کرنے کی رسائی دیں۔ نیچے سے منتخب کریں کہ کیا ترتیب دینا ہے، اور اسے کسی بھی وقت ترتیبات > AI ٹولز اور CLI میں بدلیں۔"
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "آپ جو AI ٹولز استعمال کرتے ہیں انہیں اپنے پروجیکٹس پڑھنے اور اپ ڈیٹ کرنے کی رسائی دیں۔ یہ سب بعد میں ترتیبات > AI ٹولز اور CLI میں بدلا جا سکتا ہے۔"
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "اعلیٰ آپشنز چھپائیں"
msgid "Hide Agents"
msgstr "ایجنٹس چھپائیں"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "تفصیلات چھپائیں"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "دستی طور پر پُش کیسے کریں"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "ترتیب کیسے دیں"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "{0} کیسے ترتیب دیں (براؤزر میں کھلتا ہے)"
@@ -6795,7 +6815,6 @@ msgstr "نصب کیا جا رہا ہے"
msgid "Installing {agentName}…"
msgstr "{agentName} نصب کیا جا رہا ہے…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "یہاں نصب ہوگا"
@@ -7438,7 +7457,7 @@ msgstr "MCP کی خودکار مرمت ناکام رہی"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "کوئی ایجنٹ منتخب نہیں — مہارتیں محفوظ ہ
msgid "No AI agents installed"
msgstr "کوئی AI ایجنٹ نصب نہیں"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "ابھی تک کوئی AI ٹول نہیں ملا۔ ایک انسٹال کرنے کے بعد اسے ترتیبات > AI ٹولز اور CLI سے جوڑیں۔"
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "ابھی تک کوئی AI ٹول نہیں ملا۔ ایک انسٹال کرنے کے بعد اسے ترتیبات > یہ پروجیکٹ سے جوڑیں۔"
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "کوئی AI ٹول نہیں ملا۔ اس مہارت کو استعمال کرنے کے لیے ایک نصب کریں۔"
@@ -10179,11 +10207,6 @@ msgstr "«{projectName}» کو حالیہ پروجیکٹس سے ہٹا دیا گ
msgid "Removes from"
msgstr "یہاں سے ہٹے گا"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "یہ مہارت آپ کے ایڈیٹرز سے ہٹا دیتا ہے۔"
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "ہٹایا جا رہا ہے"
@@ -10338,6 +10361,11 @@ msgstr "کیا ان نقول کو ماخذ کے سِم لنکس سے بدل دی
msgid "Replace these symlinks with independent copies?"
msgstr "کیا ان سِم لنکس کو خودمختار نقول سے بدل دیا جائے؟"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "{0} میں موجودہ OpenKnowledge اندراج کو بدل دے گا"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "اسے پلگ اِنز کے تحت اس کے اپنے صفحے پر تر
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "ان صارفین کے لیے طے شدہ مطابقت پذیری مقرر کریں جو یہ پروجیکٹ پہلی بار کھولتے ہیں۔ یہ ترتیب آپ کی ریپازٹری میں کمٹ ہوتی ہے۔"
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "{0} سیٹ اپ کریں"
@@ -11405,9 +11433,7 @@ msgstr "دستاویز کا پینل دکھائیں"
msgid "Show external URL nodes"
msgstr "بیرونی URL نوڈز دکھائیں"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "کم دکھائیں"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "مہارتیں اب مشترک ہیں۔ انہیں اپنی ٹیم کے ساتھ شیئر کرنے کے لیے .ok/skills کمٹ کریں۔"
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "مہارتیں اس مشین پر ملنے والے ہر AI ٹول میں نصب ہوتی ہیں، چاہے آپ نے اوپر کوئی بھی MCP کنکشن منتخب کیا ہو۔"
@@ -11655,11 +11679,6 @@ msgstr "skills.sh دستیاب نہیں؛ GitHub کے نتائج دکھائے ج
msgid "Skip"
msgstr "چھوڑ دیں"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "ابھی چھوڑ دیں"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "AI کو کیا کرنا چاہیے؟ (اختیاری)"
msgid "What the agent may read, write, and surface."
msgstr "ایجنٹ کیا پڑھ، لکھ اور ظاہر کر سکتا ہے۔"
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "یہ کیا بدلتا ہے"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "یہاں بنے گا: <0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "<0>{sanitized}</0> کے نام سے محفوظ ہو گا۔"
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "موجودہ OpenKnowledge اندراج کی جگہ لے لے گا"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/zh-Hans/messages.json b/packages/app/src/locales/zh-Hans/messages.json
index 22aac8fb4..9ed71ae0b 100644
--- a/packages/app/src/locales/zh-Hans/messages.json
+++ b/packages/app/src/locales/zh-Hans/messages.json
@@ -278,7 +278,6 @@
"3pU8e-": ["固定标签页"],
"3qDsdf": ["安装技巧"],
"3ri_4O": ["让您的知识库与 Open Knowledge Format 保持一致。"],
- "3s-k7L": [["0", "plural", { "other": ["再显示 ", "#", " 个工具"] }]],
"3uSzeS": ["图像可以来自任何地方 · 非图像文件必须位于工作区中"],
"3v62tp": ["skills.sh 上的热门"],
"3wiBAs": ["它应该保存在哪里?"],
@@ -556,7 +555,6 @@
["0"],
" 处有一个名字相近但完全不同的技能——那个文件夹不是这个技能。请重命名或删除其中之一以解决冲突。"
],
- "8IP5jY": ["每个选中的工具都会获得一条 OpenKnowledge MCP 条目。"],
"8Jm2EP": ["切换 1 到 6 级标题。"],
"8L0EPw": ["或者粘贴一个令牌"],
"8OGgNM": ["打开或关闭可视化编辑器的查找。"],
@@ -774,6 +772,7 @@
"C0cll5": ["已安装在 ", ["0"], "——点击可移除"],
"C0hoX6": ["始终包含"],
"C3Zc44": ["探索技能"],
+ "C4HK3A": ["添加 OpenKnowledge 的 MCP 条目和项目技能,让每个工具都能读取和更新这些文档。"],
"C62P6b": ["编辑器选区"],
"C7PKTv": ["把勾选的评论发送到对话"],
"C7waqI": [["0", "plural", { "other": ["#", " 个文档"] }]],
@@ -1334,6 +1333,7 @@
"L83DKQ": ["PATH 安装失败:", ["summary"]],
"L8fKuc": ["在你的终端中运行这个"],
"L9Y9Om": ["被链接最多的页面"],
+ "LAR9ub": ["将替换 ", ["0"], " 中已有的 OpenKnowledge 条目"],
"LBv_TR": ["把这些位置改为符号链接"],
"LBwUR6": ["切换分支超时——请尝试手动打开", ["targetNoun"], "。"],
"LCUy_N": ["会话已结束"],
@@ -1552,6 +1552,9 @@
"ODGGlj": ["文件夹名:", ["detail"]],
"OEEQKT": ["良好"],
"OETysD": ["还没有中枢页面。当页面积累起入链后,中枢页就会出现。"],
+ "OFIL46": [
+ "让你使用的 AI 工具能够读取和更新你的项目。这些之后都可以在「设置 > AI 工具与 CLI」中更改。"
+ ],
"OJ4VDl": [
"这个智能体通过 ",
["0"],
@@ -1865,7 +1868,6 @@
"T3j-7o": ["例如:同步过程中切换了项目"],
"TBfy1C": ["无法启动 Slidev。"],
"TDeOpa": ["撤销或重做你自己的编辑,协作者和智能体的编辑保持不变。"],
- "TEEPPF": ["将替换已有的 OpenKnowledge 条目"],
"TEJiKw": ["OpenKnowledge 检查了 ", ["subject"]],
"TKQ7K-": ["安装"],
"TLIO3j": ["切换斜体格式。"],
@@ -2412,6 +2414,7 @@
"。你可以从插件中安装其中任意一个。"
],
"aRI3Om": ["关闭窗口"],
+ "aSVmqV": ["尚未检测到 AI 工具。安装之后,可在「设置 > 本项目」中连接。"],
"aWlJai": ["正在启动登录流程"],
"aXkLuE": ["应用响应格式错误"],
"a_2_uB": ["无法创建 ", ["trimmed"], ":", ["0"]],
@@ -2598,7 +2601,6 @@
"dVaXnv": ["内容规则尚未加载——请稍后再试"],
"dXCi3J": ["有东西坏了"],
"dXOolp": ["正在同步你的更改"],
- "dXWMFn": ["把这个技能从你的编辑器中移除。"],
"darW93": [
[
"pageCount",
@@ -2650,6 +2652,7 @@
"ePK91l": ["编辑"],
"eQkgKV": ["已安装"],
"eQn7nI": ["正在加载配置共享"],
+ "eQsWFc": ["智能体技能"],
"eQxm13": ["更新附件位置失败——", ["detail"]],
"eRSPr9": ["打开终端"],
"eTL5y9": ["显示选区预览"],
@@ -2723,6 +2726,7 @@
"fhaxJL": ["克隆失败——响应正文为空"],
"fiMNLl": ["同步设置尚未加载——请稍后再试"],
"fj5YiP": ["从 GitHub 下载一份全新的副本。"],
+ "fmwwju": ["为每个工具添加 OpenKnowledge 的 MCP 连接。"],
"fnlBN4": ["自动批准设置尚未加载——请稍后再试"],
"fo0VXg": ["卸载"],
"fo6CCk": ["窗格 ", ["0"]],
@@ -2866,7 +2870,6 @@
"hmYd9H": [
"记录你打交道的人和公司,以及你们之间的会面。适合用来记住某人是谁、你怎么认识他们,以及上次聊了什么。"
],
- "hnlGzG": ["暂时跳过"],
"hnq8fP": ["把 ", ["label"], " 从上下文中移除"],
"hp8OtS": ["已添加"],
"hpDtUm": ["外部维基链接"],
@@ -2969,6 +2972,7 @@
"j2kt1j": [
"用你偏好的 AI 打开一个新会话。如果有选中的文本,会把它放到那里待发送,而不是直接执行。"
],
+ "j4SmVY": ["将 ", ["0"], " 连接到本项目"],
"j5CWO4": ["已是最新"],
"j9Vqvc": ["继承"],
"j9tqwO": ["不在侧边栏中"],
@@ -2993,6 +2997,7 @@
"jSO1X5": ["继承来的模板无法在这里重命名——请到拥有它的那个文件夹中重命名。"],
"jSY41h": ["“", ["docName"], "”耗时过长。请检查你的网络连接。"],
"jTLWxh": ["用内联 iframe 嵌入外部页面(文档、演示、Figma、CodeSandbox)。"],
+ "jWzDfC": ["正在检查你安装了哪些 AI 工具"],
"jZ_E5M": ["获取失败——详情请查看服务器日志。"],
"j_Z6g7": ["共享的默认值"],
"jbSG8V": ["显示「文件」(", ["sidebarShortcutLabel"], ")"],
@@ -3111,9 +3116,6 @@
"kpqWwG": ["这个", ["targetNoun"], "已从 <0>", ["branch"], "</0> 分支中移除。"],
"ktspdq": ["为存在检查或链接问题的文件着色并加标记,无需打开它们。"],
"ku0yLQ": ["页面会在搜索第一次需要它们时生成嵌入向量。"],
- "kwLcuN": [
- "让你使用的 AI 工具能够读取和更新你的项目。在下方选择要配置的项,随时可在「设置 > AI 工具与 CLI」中更改。"
- ],
"kwNS82": ["修复文件后才能使用此主题。"],
"ky9Tuc": ["编辑 ", ["0"], " 源码"],
"kzpkcn": ["无法保存那次编辑。"],
@@ -3660,6 +3662,7 @@
"tHKQ6j": ["正在重命名"],
"tKlWWY": ["表情符号"],
"tLX6LY": ["OpenKnowledge 写入了一个文档"],
+ "tLd2qa": ["尚未检测到 AI 工具。安装之后,可在「设置 > AI 工具与 CLI」中连接。"],
"tM5uzl": [
"出处、信任和生命周期系列:sources、generated、verified、status 和 stale_after。每个字段都是可选的,缺席绝不是一个错误。"
],
@@ -3884,6 +3887,7 @@
"xEb24l": ["允许我们通过邮件跟进"],
"xFQRu4": ["我本地已经有了"],
"xG4tb6": ["与“", ["docName"], "”的连接已断开。"],
+ "xGVfLh": ["继续"],
"xHGM4U": ["隐藏 HTML 预览"],
"xK291b": ["双页(奇数)"],
"xKDsGS": ["跳转到 ", ["1"], " 的第 ", ["0"], " 行"],
@@ -3964,6 +3968,9 @@
"ycIX5i": ["你已退出登录——请登录以继续同步。"],
"yckias": ["重命名文件"],
"ydzS9x": ["退出"],
+ "yeMZk8": [
+ "为每个工具添加 OpenKnowledge 的 MCP 连接,并安装 open-knowledge-discovery 技能,让你的智能体识别 OpenKnowledge 项目。"
+ ],
"yf5auF": ["按 id、别名或名称搜索规则"],
"yf5r22": [["keyName"], " 类型:", ["typeLabel"], "(继承自文件夹属性;无法在此编辑)"],
"ygCKqB": ["停止"],
diff --git a/packages/app/src/locales/zh-Hans/messages.po b/packages/app/src/locales/zh-Hans/messages.po
index db98ceeb0..e9bb8407b 100644
--- a/packages/app/src/locales/zh-Hans/messages.po
+++ b/packages/app/src/locales/zh-Hans/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, other {# 个文档均无问题。}}"
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, other {插件}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, other {再显示 # 个工具}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "添加一个受管理的块到"
msgid "Adds a managed block to {0}"
msgstr "添加一个受管理的块到 {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "添加 OpenKnowledge 的 MCP 条目和项目技能,让每个工具都能读取和更新这些文档。"
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "为每个工具添加 OpenKnowledge 的 MCP 连接,并安装 open-knowledge-discovery 技能,让你的智能体识别 OpenKnowledge 项目。"
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "为每个工具添加 OpenKnowledge 的 MCP 连接。"
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "在可用时附上遥测数据、服务器状态和运行时信息。凭据始终会被移除;如果包含文档名,它们会以明文出现(不做隐去)。"
@@ -1469,9 +1479,12 @@ msgstr "为保持 Terminal 清晰可读,已关闭智能体面板。"
msgid "Agent settings"
msgstr "智能体设置"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "智能体技能"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "智能体技能"
@@ -1488,6 +1501,8 @@ msgstr "智能体"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI 工具管理仅在 OpenKnowledge 桌面应用中可用。"
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "检查技能"
msgid "Checking source"
msgstr "正在检查来源"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "正在检查你安装了哪些 AI 工具"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "正在检查你的 GitHub 连接"
@@ -2925,12 +2944,15 @@ msgstr "冲突文件:{0}。"
msgid "Conflicts"
msgstr "冲突"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "连接"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "将 {0} 连接到本项目"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "连接一个 GitHub 账户,以便浏览和同步你的仓库。"
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "把这个项目连接到 GitHub 才能共享。"
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "连接 AI 工具"
@@ -3063,7 +3084,6 @@ msgstr "内容规则"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "内容规则尚未加载——请稍后再试"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "上下文开销"
@@ -3076,6 +3096,11 @@ msgstr "上下文窗口:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "上下文窗口:已使用 {percent}%,{usedLabel}/{sizeLabel} 个 token"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "继续"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "在浏览器中继续"
@@ -4661,10 +4686,6 @@ msgstr "例如:粘贴一个很大的表格后编辑器卡住了"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "每个包都会用现成的文件夹和模板为你的项目搭好骨架。"
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "每个选中的工具都会获得一条 OpenKnowledge MCP 条目。"
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "输入有效的 URL,例如 https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "输入你的隧道所服务的公开来源,例如 https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "字段"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "让你使用的 AI 工具能够读取和更新你的项目。勾选即刻生效;取消勾选则移除。"
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "让你使用的 AI 工具能够读取和更新你的项目。在下方选择要配置的项,随时可在「设置 > AI 工具与 CLI」中更改。"
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "让你使用的 AI 工具能够读取和更新你的项目。这些之后都可以在「设置 > AI 工具与 CLI」中更改。"
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "隐藏高级选项"
msgid "Hide Agents"
msgstr "隐藏「智能体」"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "隐藏详情"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "如何手动推送"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "如何设置"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "如何设置 {0}(在浏览器中打开)"
@@ -6795,7 +6815,6 @@ msgstr "正在安装"
msgid "Installing {agentName}…"
msgstr "正在安装 {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "安装到"
@@ -7438,7 +7457,7 @@ msgstr "MCP 自动修复失败"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "未选择任何智能体——技能已保存,但没有安装到任何
msgid "No AI agents installed"
msgstr "未安装任何 AI 智能体"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "尚未检测到 AI 工具。安装之后,可在「设置 > AI 工具与 CLI」中连接。"
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "尚未检测到 AI 工具。安装之后,可在「设置 > 本项目」中连接。"
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "未检测到 AI 工具。请先安装一个再使用此技能。"
@@ -10179,11 +10207,6 @@ msgstr "已把“{projectName}”从最近的项目中移除。它的文件夹
msgid "Removes from"
msgstr "从以下位置移除"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "把这个技能从你的编辑器中移除。"
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "正在移除"
@@ -10338,6 +10361,11 @@ msgstr "要把这些副本替换为指向来源的符号链接吗?"
msgid "Replace these symlinks with independent copies?"
msgstr "要把这些符号链接替换为独立副本吗?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "将替换 {0} 中已有的 OpenKnowledge 条目"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "在「插件」下它自己的页面中进行设置。"
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "为首次打开这个项目的用户设置同步默认值。这项设置会提交到你的仓库。"
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "设置 {0}"
@@ -11405,9 +11433,7 @@ msgstr "显示文档面板"
msgid "Show external URL nodes"
msgstr "显示外部 URL 节点"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "显示更少"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "技能现在已共享。提交 .ok/skills 即可分享给你的团队。"
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "技能会安装到本机检测到的每个 AI 工具,与上面选择的 MCP 连接无关。"
@@ -11655,11 +11679,6 @@ msgstr "skills.sh 不可用;改为显示 GitHub 结果(没有安装量排名
msgid "Skip"
msgstr "跳过"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "暂时跳过"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "希望 AI 做什么?(可选)"
msgid "What the agent may read, write, and surface."
msgstr "智能体可以读取、写入和呈现的内容。"
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "这会改变什么"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "将创建在:<0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "将保存为 <0>{sanitized}</0>。"
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "将替换已有的 OpenKnowledge 条目"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/app/src/locales/zh-Hant/messages.json b/packages/app/src/locales/zh-Hant/messages.json
index 3ec258ca8..5fe759545 100644
--- a/packages/app/src/locales/zh-Hant/messages.json
+++ b/packages/app/src/locales/zh-Hant/messages.json
@@ -282,7 +282,6 @@
"3pU8e-": ["釘選分頁"],
"3qDsdf": ["安裝技巧"],
"3ri_4O": ["讓您的知識庫與 Open Knowledge Format 保持一致。"],
- "3s-k7L": [["0", "plural", { "other": ["再顯示 ", "#", " 個工具"] }]],
"3uSzeS": ["圖片可以來自任何地方 · 非圖片檔案必須位於工作區中"],
"3v62tp": ["skills.sh 上的熱門項目"],
"3wiBAs": ["它應該放在哪裡?"],
@@ -560,7 +559,6 @@
["0"],
" 有一個名稱相近但完全不同的技能——那個資料夾不是這個技能。請重新命名或刪除其中之一來解決。"
],
- "8IP5jY": ["每個選取的工具都會得到一筆 OpenKnowledge MCP 項目。"],
"8Jm2EP": ["切換 1 到 6 級標題。"],
"8L0EPw": ["或者貼上一個權杖"],
"8OGgNM": ["開啟或關閉視覺化編輯器的尋找。"],
@@ -778,6 +776,7 @@
"C0cll5": ["已安裝在 ", ["0"], "——點一下可移除"],
"C0hoX6": ["一律包含"],
"C3Zc44": ["探索技能"],
+ "C4HK3A": ["新增 OpenKnowledge 的 MCP 項目與專案技能,讓每個工具都能讀取並更新這些文件。"],
"C62P6b": ["編輯器選取範圍"],
"C7PKTv": ["將勾選的註解傳送到對話"],
"C7waqI": [["0", "plural", { "other": ["#", " 個文件"] }]],
@@ -1350,6 +1349,7 @@
"L83DKQ": ["PATH 安裝失敗:", ["summary"]],
"L8fKuc": ["在你的終端機中執行這個"],
"L9Y9Om": ["被連結最多的頁面"],
+ "LAR9ub": ["將取代 ", ["0"], " 中既有的 OpenKnowledge 項目"],
"LBv_TR": ["把這些位置改成符號連結"],
"LBwUR6": ["切換分支逾時——請試著手動開啟", ["targetNoun"], "。"],
"LCUy_N": ["工作階段已結束"],
@@ -1574,6 +1574,9 @@
"ODGGlj": ["資料夾名稱:", ["detail"]],
"OEEQKT": ["良好"],
"OETysD": ["還沒有樞紐頁面。等頁面累積起連入的關係圖連結,樞紐頁就會出現。"],
+ "OFIL46": [
+ "讓你使用的 AI 工具能夠讀取並更新你的專案。這些之後都可以在「設定 > AI 工具與 CLI」中更改。"
+ ],
"OJ4VDl": [
"這個代理是透過 ",
["0"],
@@ -1889,7 +1892,6 @@
"T3j-7o": ["例如:同步進行到一半時切換了專案"],
"TBfy1C": ["無法啟動 Slidev。"],
"TDeOpa": ["復原或重做你自己的編輯,協作者和代理的編輯維持不變。"],
- "TEEPPF": ["會取代既有的 OpenKnowledge 項目"],
"TEJiKw": ["OpenKnowledge 檢查了 ", ["subject"]],
"TKQ7K-": ["安裝"],
"TLIO3j": ["切換斜體格式。"],
@@ -2436,6 +2438,7 @@
"。你可以從外掛中安裝其中任何一個。"
],
"aRI3Om": ["關閉視窗"],
+ "aSVmqV": ["尚未偵測到 AI 工具。安裝之後,可在「設定 > 這個專案」中連接。"],
"aWlJai": ["正在啟動登入流程"],
"aXkLuE": ["套用回應格式錯誤"],
"a_2_uB": ["無法建立 ", ["trimmed"], ":", ["0"]],
@@ -2622,7 +2625,6 @@
"dVaXnv": ["內容規則尚未載入——請稍候再試"],
"dXCi3J": ["有東西壞了"],
"dXOolp": ["正在同步你的變更"],
- "dXWMFn": ["把這個技能從你的編輯器中移除。"],
"darW93": [
[
"pageCount",
@@ -2674,6 +2676,7 @@
"ePK91l": ["編輯"],
"eQkgKV": ["已安裝"],
"eQn7nI": ["正在載入設定分享"],
+ "eQsWFc": ["代理技能"],
"eQxm13": ["更新附件位置失敗——", ["detail"]],
"eRSPr9": ["開啟終端機"],
"eTL5y9": ["顯示選取範圍預覽"],
@@ -2747,6 +2750,7 @@
"fhaxJL": ["複製失敗——回應內容為空"],
"fiMNLl": ["同步設定尚未載入——請稍候再試"],
"fj5YiP": ["從 GitHub 下載一份全新的副本。"],
+ "fmwwju": ["為每個工具新增 OpenKnowledge 的 MCP 連接。"],
"fnlBN4": ["自動核准設定尚未載入——請稍候再試"],
"fo0VXg": ["解除安裝"],
"fo6CCk": ["窗格 ", ["0"]],
@@ -2894,7 +2898,6 @@
"hmYd9H": [
"記錄你往來的人和公司,以及你們之間的會面。適合用來記住某人是誰、你怎麼認識他們,還有上次聊了什麼。"
],
- "hnlGzG": ["先略過"],
"hnq8fP": ["把 ", ["label"], " 從脈絡中移除"],
"hp8OtS": ["已新增"],
"hpDtUm": ["外部維基連結"],
@@ -2997,6 +3000,7 @@
"j2kt1j": [
"用你偏好的 AI 開啟一個新的工作階段。如果有選取文字,會把它放到那裡待送出,而不是直接執行。"
],
+ "j4SmVY": ["將 ", ["0"], " 連接到這個專案"],
"j5CWO4": ["已是最新"],
"j9Vqvc": ["繼承"],
"j9tqwO": ["不在側邊欄中"],
@@ -3021,6 +3025,7 @@
"jSO1X5": ["繼承來的範本無法在這裡重新命名——請到擁有它的那個資料夾中重新命名。"],
"jSY41h": ["「", ["docName"], "」耗時過久。請檢查你的網路連線。"],
"jTLWxh": ["用行內 iframe 嵌入外部頁面(說明文件、示範、Figma、CodeSandbox)。"],
+ "jWzDfC": ["正在檢查你安裝了哪些 AI 工具"],
"jZ_E5M": ["取得失敗——詳情請看伺服器記錄。"],
"j_Z6g7": ["分享的預設值"],
"jbSG8V": ["顯示「檔案」(", ["sidebarShortcutLabel"], ")"],
@@ -3139,9 +3144,6 @@
"kpqWwG": ["這個", ["targetNoun"], "已經從 <0>", ["branch"], "</0> 分支移除了。"],
"ktspdq": ["為有檢查或連結問題的檔案上色並加上標記,不必開啟它們。"],
"ku0yLQ": ["頁面會在搜尋第一次需要它們的時候產生嵌入向量。"],
- "kwLcuN": [
- "讓你用的 AI 工具能夠讀取和更新你的專案。在下面挑要設定的項目,之後隨時可以在「設定 > AI 工具與 CLI」中變更。"
- ],
"kwNS82": ["修正檔案後才能使用這個主題。"],
"ky9Tuc": ["編輯 ", ["0"], " 原始碼"],
"kzpkcn": ["無法儲存那次編輯。"],
@@ -3690,6 +3692,7 @@
"tHKQ6j": ["正在重新命名"],
"tKlWWY": ["表情符號"],
"tLX6LY": ["OpenKnowledge 寫入了一個文件"],
+ "tLd2qa": ["尚未偵測到 AI 工具。安裝之後,可在「設定 > AI 工具與 CLI」中連接。"],
"tM5uzl": [
"來源、信任與生命週期系列:sources、generated、verified、status 和 stale_after。每個欄位都是可選的,缺席絕不是一個錯誤。"
],
@@ -3916,6 +3919,7 @@
"xEb24l": ["允許我們用電子郵件跟進"],
"xFQRu4": ["我本機已經有了"],
"xG4tb6": ["與「", ["docName"], "」的連線已中斷。"],
+ "xGVfLh": ["繼續"],
"xHGM4U": ["隱藏 HTML 預覽"],
"xK291b": ["雙頁(奇數)"],
"xKDsGS": ["移至 ", ["1"], " 的第 ", ["0"], " 行"],
@@ -3996,6 +4000,9 @@
"ycIX5i": ["你已登出——請登入以繼續同步。"],
"yckias": ["重新命名檔案"],
"ydzS9x": ["結束"],
+ "yeMZk8": [
+ "為每個工具新增 OpenKnowledge 的 MCP 連接,並安裝 open-knowledge-discovery 技能,讓你的智慧代理辨識 OpenKnowledge 專案。"
+ ],
"yf5auF": ["依 id、別名或名稱搜尋規則"],
"yf5r22": [["keyName"], " 類型:", ["typeLabel"], "(繼承自資料夾屬性;無法在此編輯)"],
"ygCKqB": ["停止"],
diff --git a/packages/app/src/locales/zh-Hant/messages.po b/packages/app/src/locales/zh-Hant/messages.po
index a4149befa..314b2b707 100644
--- a/packages/app/src/locales/zh-Hant/messages.po
+++ b/packages/app/src/locales/zh-Hant/messages.po
@@ -278,11 +278,6 @@ msgstr "{0, plural, other {# 份文件都沒有問題。}}"
msgid "{0, plural, one {Plugin} other {Plugins}}"
msgstr "{0, plural, other {外掛程式}}"
-#. placeholder {0}: editorList.hiddenCount
-#: src/components/McpConsentDialogBody.tsx
-msgid "{0, plural, one {Show # more tool} other {Show # more tools}}"
-msgstr "{0, plural, other {再顯示 # 個工具}}"
-
#. placeholder {0}: skillRows.length
#: src/components/CreatedItemsList.tsx
msgid "{0, plural, one {Skill} other {Skills}}"
@@ -1414,6 +1409,21 @@ msgstr "加入一個受管理的區塊到"
msgid "Adds a managed block to {0}"
msgstr "加入一個受管理的區塊到 {0}"
+#. Subtext under the create-project AI-tools checkbox
+#: src/components/CreateProjectDialog.tsx
+msgid "Adds an OpenKnowledge MCP entry and the project skill, so each tool can read and update these notes."
+msgstr "新增 OpenKnowledge 的 MCP 項目與專案技能,讓每個工具都能讀取並更新這些文件。"
+
+#. Subtext under the single AI-tools checkbox when a skill is also installed
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each, plus the open-knowledge-discovery skill so your agents recognize OpenKnowledge projects."
+msgstr "為每個工具新增 OpenKnowledge 的 MCP 連接,並安裝 open-knowledge-discovery 技能,讓你的智慧代理辨識 OpenKnowledge 專案。"
+
+#. Subtext under the single AI-tools checkbox when only MCP is wired
+#: src/components/McpConsentDialogBody.tsx
+msgid "Adds OpenKnowledge's MCP connection to each."
+msgstr "為每個工具新增 OpenKnowledge 的 MCP 連接。"
+
#: src/components/ReportBugDialogBody.tsx
msgid "Adds telemetry, server state, and runtime info when available. Credentials are always removed; document names, if included, appear in cleartext (not redacted)."
msgstr "在可取得時附上遙測資料、伺服器狀態和執行階段資訊。憑證一律會移除;如果包含文件名稱,它們會以明文出現(不做遮蔽)。"
@@ -1469,9 +1479,12 @@ msgstr "為保持 Terminal 清晰可讀,已關閉智慧代理面板。"
msgid "Agent settings"
msgstr "代理設定"
-#. Group label above the skill rows in Settings → AI tools & CLI
-#. Section label above the skill checkboxes in the first-launch dialog
+#. Heading above the skill install destinations in the disclosure
#: src/components/McpConsentDialogBody.tsx
+msgid "Agent Skill"
+msgstr "代理技能"
+
+#. Group label above the skill rows in Settings → AI tools & CLI
#: src/components/settings/AiToolsSection.tsx
msgid "Agent Skills"
msgstr "代理技能"
@@ -1488,6 +1501,8 @@ msgstr "代理"
msgid "AI tool management is only available in the OpenKnowledge desktop app."
msgstr "AI 工具管理只在 OpenKnowledge 桌面應用程式中提供。"
+#. Section label above the AI-tools checkbox in the first-launch dialog
+#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/settings/SettingsDialogShell.tsx
msgid "AI tools"
@@ -2375,6 +2390,10 @@ msgstr "檢查技能"
msgid "Checking source"
msgstr "正在檢查來源"
+#: src/components/CreateProjectDialog.tsx
+msgid "Checking which AI tools you have"
+msgstr "正在檢查你安裝了哪些 AI 工具"
+
#: src/components/settings/AccountSection.tsx
msgid "Checking your GitHub connection"
msgstr "正在檢查你的 GitHub 連線"
@@ -2925,12 +2944,15 @@ msgstr "衝突的檔案:{0}。"
msgid "Conflicts"
msgstr "衝突"
-#. Primary button that writes MCP config for the selected AI tools
#: src/components/AuthModal.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Connect"
msgstr "連接"
+#. placeholder {0}: formatToolList( status.editors.map((id) => EDITOR_LABELS[id]), i18n.locale, )
+#: src/components/CreateProjectDialog.tsx
+msgid "Connect {0} to this project"
+msgstr "將 {0} 連接到這個專案"
+
#: src/components/settings/AccountSection.tsx
msgid "Connect a GitHub account to browse and sync your repositories."
msgstr "連接一個 GitHub 帳戶,就能瀏覽並同步你的儲存庫。"
@@ -2964,7 +2986,6 @@ msgid "Connect this project to GitHub to share."
msgstr "把這個專案連接到 GitHub 才能分享。"
#: src/components/ConsentDialogBody.tsx
-#: src/components/CreateProjectDialog.tsx
msgid "Connect to AI tools"
msgstr "連接 AI 工具"
@@ -3063,7 +3084,6 @@ msgstr "內容規則"
msgid "Content rules not yet loaded — try again in a moment"
msgstr "內容規則尚未載入——請稍候再試"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Context cost"
msgstr "上下文開銷"
@@ -3076,6 +3096,11 @@ msgstr "脈絡視窗:"
msgid "Context window: {percent}% used, {usedLabel} of {sizeLabel} tokens"
msgstr "脈絡視窗:已使用 {percent}%,{usedLabel}/{sizeLabel} 個 token"
+#. Primary button that applies the first-launch setup choices
+#: src/components/McpConsentDialogBody.tsx
+msgid "Continue"
+msgstr "繼續"
+
#: src/components/AuthModal.tsx
msgid "Continue in browser"
msgstr "在瀏覽器中繼續"
@@ -4661,10 +4686,6 @@ msgstr "例如:貼上一個很大的表格之後編輯器卡住了"
msgid "Each pack scaffolds your project with ready-made folders and templates."
msgstr "每個套件都會用現成的資料夾和範本幫你的專案搭好骨架。"
-#: src/components/CreateProjectDialog.tsx
-msgid "Each selected tool gets an OpenKnowledge MCP entry."
-msgstr "每個選取的工具都會得到一筆 OpenKnowledge MCP 項目。"
-
#: src/components/AppMenubar.tsx
#: src/components/settings/LintingSection.tsx
#: src/components/SkillsTree.tsx
@@ -5021,7 +5042,6 @@ msgstr "輸入有效的 URL,例如 https://notes.example.com"
msgid "Enter the public origin your tunnel serves, for example https://notes.example.com"
msgstr "輸入你的通道所提供服務的公開來源,例如 https://notes.example.com"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "Entry"
@@ -5531,7 +5551,6 @@ msgstr "欄位"
#: src/components/AppMenubar.tsx
#: src/components/CommandPalette.tsx
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
#: src/components/share-metadata-rows.tsx
@@ -5981,8 +6000,8 @@ msgid "Give the AI tools you use access to read and update your projects. Checki
msgstr "讓你用的 AI 工具能夠讀取和更新你的專案。勾選就會立刻設定好;取消勾選則會移除。"
#: src/components/McpConsentDialogBody.tsx
-msgid "Give the AI tools you use access to read and update your projects. Pick what to set up below, and change it anytime in Settings > AI tools & CLI."
-msgstr "讓你用的 AI 工具能夠讀取和更新你的專案。在下面挑要設定的項目,之後隨時可以在「設定 > AI 工具與 CLI」中變更。"
+msgid "Give the AI tools you use access to read and update your projects. You can change any of this later in Settings > AI tools & CLI."
+msgstr "讓你使用的 AI 工具能夠讀取並更新你的專案。這些之後都可以在「設定 > AI 工具與 CLI」中更改。"
#: src/components/settings/LintingSection.tsx
msgid "Glob patterns this schema applies to"
@@ -6164,7 +6183,11 @@ msgstr "隱藏進階選項"
msgid "Hide Agents"
msgstr "隱藏「代理」"
+#. Collapses the list of files the setup writes to
+#. Collapses the list of project files the AI-tool setup writes
#: src/components/acp/ThreadView.tsx
+#: src/components/CreateProjectDialog.tsx
+#: src/components/McpConsentDialogBody.tsx
msgid "Hide details"
msgstr "隱藏詳細資料"
@@ -6306,14 +6329,11 @@ msgid "How to push manually"
msgstr "如何手動推送"
#. Link on an undetected tool row to its OpenKnowledge setup guide
-#. Link on an undetected tool row to its OpenKnowledge setup guide
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up"
msgstr "如何設定"
#. placeholder {0}: editor.label
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "How to set up {0} (opens in browser)"
msgstr "如何設定 {0}(在瀏覽器中開啟)"
@@ -6795,7 +6815,6 @@ msgstr "正在安裝"
msgid "Installing {agentName}…"
msgstr "正在安裝 {agentName}…"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/SkillInstallConfirmDialog.tsx
msgid "Installs to"
msgstr "安裝到"
@@ -7438,7 +7457,7 @@ msgstr "MCP 自動修復失敗"
#. Group label above the per-editor MCP list in Settings → AI tools & CLI — each row wires OpenKnowledge's MCP server into that tool
#. Group label above the per-editor project MCP list in Settings → This project → AI tools
-#. Section label above the editor checkbox list in the first-launch dialog — each row wires OpenKnowledge's MCP server into that tool
+#. Heading above the per-tool MCP config paths in the disclosure
#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
@@ -8046,6 +8065,15 @@ msgstr "沒有選取任何代理——技能已儲存,但沒有安裝到任何
msgid "No AI agents installed"
msgstr "沒有安裝任何 AI 代理"
+#. Shown in place of the AI-tools checkbox when no AI tool was detected
+#: src/components/McpConsentDialogBody.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > AI tools & CLI."
+msgstr "尚未偵測到 AI 工具。安裝之後,可在「設定 > AI 工具與 CLI」中連接。"
+
+#: src/components/CreateProjectDialog.tsx
+msgid "No AI tools detected yet. Once you install one, connect it from Settings > This project."
+msgstr "尚未偵測到 AI 工具。安裝之後,可在「設定 > 這個專案」中連接。"
+
#: src/components/SkillConsentRow.tsx
msgid "No AI tools detected. Install one to use this skill."
msgstr "未偵測到 AI 工具。請先安裝一個再使用此技能。"
@@ -10179,11 +10207,6 @@ msgstr "已把「{projectName}」從最近的專案中移除。它的資料夾
msgid "Removes from"
msgstr "從以下位置移除"
-#. Warning shown when the user unchecks an already-installed skill
-#: src/components/McpConsentDialogBody.tsx
-msgid "Removes this skill from your editors."
-msgstr "把這個技能從你的編輯器中移除。"
-
#: src/components/CreateProjectDialog.tsx
msgid "Removing"
msgstr "正在移除"
@@ -10338,6 +10361,11 @@ msgstr "要把這些副本換成指向來源的符號連結嗎?"
msgid "Replace these symlinks with independent copies?"
msgstr "要把這些符號連結換成獨立的副本嗎?"
+#. placeholder {0}: formatToolList( replacing.map((e) => e.label), i18n.locale, )
+#: src/components/McpConsentDialogBody.tsx
+msgid "Replaces the existing OpenKnowledge entry in {0}"
+msgstr "將取代 {0} 中既有的 OpenKnowledge 項目"
+
#: src/components/BugReportHistory.tsx
#: src/components/command-palette-commands.ts
#: src/components/HelpPopover.tsx
@@ -11213,7 +11241,7 @@ msgstr "在「外掛」底下它自己的頁面中設定。"
msgid "Set the sync default for users opening this project for the first time. This setting is committed to your repository."
msgstr "為第一次開啟這個專案的使用者設定同步預設值。這項設定會提交到你的儲存庫。"
-#. placeholder {0}: skill.name
+#. placeholder {0}: formatToolList( editors.map((e) => e.label), i18n.locale, )
#: src/components/McpConsentDialogBody.tsx
msgid "Set up {0}"
msgstr "設定 {0}"
@@ -11405,9 +11433,7 @@ msgstr "顯示文件面板"
msgid "Show external URL nodes"
msgstr "顯示外部 URL 節點"
-#. Collapses the undetected AI tools back behind the toggle
#: src/components/LinksPanel.tsx
-#: src/components/McpConsentDialogBody.tsx
msgid "Show fewer"
msgstr "顯示少一點"
@@ -11626,8 +11652,6 @@ msgid "Skills are now shared. Commit .ok/skills to share them with your team."
msgstr "技能現在已分享。提交 .ok/skills 就能分享給你的團隊。"
#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#. Tells the user a skill installs to every detected AI tool, independent of the per-tool MCP connections listed above
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
msgid "Skills install to every AI tool detected on this machine, independent of the MCP connections you chose above."
msgstr "技能會安裝到本機偵測到的每個 AI 工具,與上面選擇的 MCP 連線無關。"
@@ -11655,11 +11679,6 @@ msgstr "skills.sh 無法使用;改為顯示 GitHub 結果(沒有安裝次數
msgid "Skip"
msgstr "略過"
-#. Secondary button — dismisses the dialog without wiring any tools
-#: src/components/McpConsentDialogBody.tsx
-msgid "Skip for now"
-msgstr "先略過"
-
#. placeholder {0}: item.displayName
#: src/components/acp/ThreadView.tsx
msgid "Skipped downloading {0}."
@@ -13993,11 +14012,13 @@ msgstr "希望 AI 做什麼?(選填)"
msgid "What the agent may read, write, and surface."
msgstr "代理可以讀取、寫入和呈現的內容。"
+#. Expands the list of files the setup writes to
+#. Expands the list of project files the AI-tool setup writes
+#: src/components/CreateProjectDialog.tsx
#: src/components/McpConsentDialogBody.tsx
msgid "What this changes"
msgstr "這會改變什麼"
-#: src/components/McpConsentDialogBody.tsx
#: src/components/settings/AiToolsSection.tsx
#: src/components/settings/ProjectAiToolsSection.tsx
msgid "What this checkbox changes"
@@ -14091,11 +14112,6 @@ msgstr "會建立在:<0>{targetPreview}</0>"
msgid "Will be saved as <0>{sanitized}</0>."
msgstr "會儲存成 <0>{sanitized}</0>。"
-#. Disclosure that Add will overwrite the tool's existing OpenKnowledge MCP entry
-#: src/components/McpConsentDialogBody.tsx
-msgid "Will replace existing OpenKnowledge entry"
-msgstr "會取代既有的 OpenKnowledge 項目"
-
#: src/components/AppMenubar.tsx
#: src/lib/native-menu-catalog.ts
msgid "Window"
diff --git a/packages/cli/src/integrations/write-project-skill.ts b/packages/cli/src/integrations/write-project-skill.ts
index 50a251aa9..de0420163 100644
--- a/packages/cli/src/integrations/write-project-skill.ts
+++ b/packages/cli/src/integrations/write-project-skill.ts
@@ -16,6 +16,7 @@ import {
statSync,
} from 'node:fs';
import { dirname, isAbsolute, relative, resolve, sep } from 'node:path';
+import { USER_MCP_GATED_EDITOR_IDS } from '@inkeep/open-knowledge-core';
import { resolveBundledSkillDir } from '@inkeep/open-knowledge-server';
import { type ParseError, parse as parseJsonc } from 'jsonc-parser';
import type { EditorId, EditorMcpTarget } from '../commands/editors.ts';
@@ -170,13 +171,17 @@ function configHasOpenKnowledgeEntry(
* Copilot loads OK's runtime skill from the project but its MCP registration
* from user-global config. Shared project MCP wiring belongs to other hosts and
* cannot satisfy Copilot's prerequisite.
+ *
+ * The gated set lives in core so UI that names which tools a project setup
+ * covers reads the same list this write path enforces — otherwise a picker can
+ * promise a skill install that lands as `skipped-prerequisite`.
*/
function isProjectSkillPrerequisiteMet(
target: EditorMcpTarget,
cwd: string,
options: ProjectSkillWriteOptions = {},
): boolean {
- if (target.id !== 'copilot') return true;
+ if (!USER_MCP_GATED_EDITOR_IDS.includes(target.id)) return true;
try {
return configHasOpenKnowledgeEntry(target, cwd, target.configPath(cwd, options.home));
diff --git a/packages/core/src/constants/editors.ts b/packages/core/src/constants/editors.ts
index 6297b35bd..64ac3c3fc 100644
--- a/packages/core/src/constants/editors.ts
+++ b/packages/core/src/constants/editors.ts
@@ -317,3 +317,39 @@ export const EDITOR_PROJECT_CONFIG_PATH = {
// "project-configured" — same as OpenClaw.
hermes: null,
} as const satisfies Record<EditorId, string | null>;
+
+/**
+ * Editors whose PROJECT skill only loads once their USER-GLOBAL MCP entry
+ * exists. Copilot reads the skill from the project (`.github/skills`) but takes
+ * its OpenKnowledge registration from user-global config, and project MCP wiring
+ * belongs to other hosts — so writing the skill before the global entry is there
+ * produces a file Copilot never loads.
+ *
+ * The write path enforces this in `isProjectSkillPrerequisiteMet`; UI that names
+ * which tools a project setup covers reads it through
+ * {@link receivesProjectIntegrationWrite}. One list, so a screen can't claim a
+ * write the writer will skip.
+ *
+ * Named `*_EDITOR_IDS` like every other filtered id list here; the `EDITOR_*`
+ * prefix in this file is reserved for `Record<EditorId, T>` maps.
+ */
+export const USER_MCP_GATED_EDITOR_IDS: readonly EditorId[] = ['copilot'];
+
+/**
+ * Whether setting up a project for `id` actually writes something, given
+ * whether that editor's user-global OpenKnowledge MCP entry is already
+ * installed. This is the predicate a project-scoped picker owes the user: it
+ * answers "will ticking this produce a file?", not "does this editor have a
+ * surface in principle".
+ *
+ * A project MCP config always lands. A project skill lands unless the editor
+ * gates it on a user-global entry that isn't there yet.
+ */
+export function receivesProjectIntegrationWrite(
+ id: EditorId,
+ opts: { userMcpEntryInstalled: boolean },
+): boolean {
+ if (EDITOR_PROJECT_CONFIG_PATH[id] !== null) return true;
+ if (EDITOR_PROJECT_SKILL_ROOT[id] === null) return false;
+ return !USER_MCP_GATED_EDITOR_IDS.includes(id) || opts.userMcpEntryInstalled;
+}
diff --git a/packages/core/src/desktop-bridge.ts b/packages/core/src/desktop-bridge.ts
index 8aa981646..655702e06 100644
--- a/packages/core/src/desktop-bridge.ts
+++ b/packages/core/src/desktop-bridge.ts
@@ -522,14 +522,15 @@ export type OkMcpWiringEditorId = EditorId;
/**
* Payload delivered to `mcpWiring.onShow` subscribers on first-launch MCP
- * consent. Every editor in `ALL_EDITOR_IDS` appears; `detected: true`
- * preselects the checkbox in `<McpConsentDialog>`. `willReplace: true`
- * signals that the editor has an existing OK-managed entry that Add would
- * overwrite — surfaced per-row so long-time CLI users aren't surprised to
- * find their pre-existing entry stomped. `pathInstall` drives the dialog's
- * shell-PATH toggle row: `shellDetected: false` hides the row;
- * `alreadyInstalled: true` renders it informational; `rcFilesToTouch`
- * names the tildified shell files a grant would edit.
+ * consent. Every editor in `ALL_EDITOR_IDS` appears, but only `detected: true`
+ * ones are in `<McpConsentDialog>`'s write set — the rest have no config to
+ * wire and are not shown. `willReplace: true` signals an existing OK-managed
+ * entry the setup would overwrite, which the dialog surfaces next to its
+ * checkbox so long-time CLI users aren't surprised to find their pre-existing
+ * entry stomped. `pathInstall` drives the dialog's shell-PATH toggle row:
+ * `shellDetected: false` hides the row; `alreadyInstalled: true` renders it
+ * informational; `rcFilesToTouch` names the tildified shell files a grant
+ * would edit.
*/
export interface OkMcpWiringShowPayload {
readonly detectedEditors: readonly {
@@ -547,24 +548,14 @@ export interface OkMcpWiringShowPayload {
readonly rcFilesToTouch: readonly string[];
readonly alreadyInstalled: boolean;
};
- /** Per-bundle opt-in rows for the two user-global skills. Empty ⇒ no skill
- * decision solicited. `alreadyInstalled: true` renders the row pre-checked
- * as an existing install the user can uncheck to remove. The disclosure
- * fields (description, size, hosts, paths) let the row show what the skill
- * is, costs, and where it would land — the same information Settings shows,
- * so first-launch consent is as informed. `hosts` empty ⇒ nowhere to land,
- * and the row disables its checkbox. */
+ /** The user-global skill bundles onboarding sets up alongside the MCP wiring.
+ * Empty ⇒ no skill decision solicited. `paths` lists every destination the
+ * install writes to, computed from the installer's own iteration set and
+ * gates, so the dialog's disclosure can never advertise a copy that will not
+ * be made. */
readonly globalSkills: readonly {
readonly id: string;
readonly name: string;
- readonly alreadyInstalled: boolean;
- /** The skill's own frontmatter description. */
- readonly description: string;
- /** Three-tier context cost; absent when the packaged bundle is unreadable. */
- readonly size?: SkillCostTiers;
- /** Resolved reach: static editor ids plus declared custom-root paths. */
- readonly hosts: readonly string[];
- /** Destination paths a confirmed install writes (or an uninstall removes). */
readonly paths: readonly string[];
}[];
}
@@ -575,9 +566,12 @@ export interface OkMcpWiringShowPayload {
* granted); `false` → record declined, touch no rc file; absent → no PATH
* decision was solicited (row hidden or informational).
*
- * `skills` — bundle ids the user left checked. Present ⇒ a skill decision was
- * solicited; every offered bundle not in the list is recorded declined (and
- * removed if already installed).
+ * `skills` — bundle ids the user left checked. An ARRAY (even empty) ⇒ a
+ * decision was made, and every offered bundle not in the list is recorded
+ * declined (and removed if already installed). Absent ⇒ no decision was made:
+ * the skills leg is skipped, so nothing is written and nothing installed is
+ * torn down. Onboarding sends the absent form when the user declines setup,
+ * because declining must never uninstall an existing bundle.
*/
export interface OkMcpWiringConfirmRequest {
readonly editorIds: readonly OkMcpWiringEditorId[];
diff --git a/packages/core/src/index.ts b/packages/core/src/index.ts
index 79c9a4cee..43781374d 100644
--- a/packages/core/src/index.ts
+++ b/packages/core/src/index.ts
@@ -351,7 +351,9 @@ export {
PROJECT_SKILL_EDITOR_IDS,
PROJECT_SKILL_PROJECTION_IGNORE_PATHS,
RESERVED_PROJECT_SKILL_NAME,
+ receivesProjectIntegrationWrite,
skillRootActivationPath,
+ USER_MCP_GATED_EDITOR_IDS,
USER_SKILL_HOSTS,
} from './constants/editors.ts';
export {
diff --git a/packages/core/src/skill-targets/editor-single-source.test.ts b/packages/core/src/skill-targets/editor-single-source.test.ts
index eceb8b82f..318781e03 100644
--- a/packages/core/src/skill-targets/editor-single-source.test.ts
+++ b/packages/core/src/skill-targets/editor-single-source.test.ts
@@ -17,10 +17,13 @@
import { describe, expect, test } from 'vitest';
import {
ALL_EDITOR_IDS,
+ EDITOR_PROJECT_CONFIG_PATH,
EDITOR_PROJECT_SKILL_ROOT,
EDITOR_USER_SKILL_ROOT,
HOSTS_WITH_USER_SKILL_DIR,
PROJECT_SKILL_EDITOR_IDS,
+ receivesProjectIntegrationWrite,
+ USER_MCP_GATED_EDITOR_IDS,
USER_SKILL_HOSTS,
} from '../constants/editors.ts';
import { SkillTargetEditorSchema } from './schema.ts';
@@ -95,3 +98,53 @@ describe('project-skill editor-id single source', () => {
expect(SkillTargetEditorSchema.options).not.toContain('claude-desktop');
});
});
+
+describe('receivesProjectIntegrationWrite', () => {
+ // Surface membership is necessary but not sufficient: this predicate is what
+ // a project-scoped picker owes the user, because it answers "will ticking
+ // this produce a file?" rather than "could it in principle?".
+ const installed = { userMcpEntryInstalled: true };
+ const notInstalled = { userMcpEntryInstalled: false };
+
+ test('an editor with a project MCP config always writes, whatever the global state', () => {
+ for (const id of ['claude', 'cursor', 'codex', 'opencode', 'pi'] as const) {
+ expect(EDITOR_PROJECT_CONFIG_PATH[id]).not.toBeNull();
+ expect(receivesProjectIntegrationWrite(id, installed)).toBe(true);
+ expect(receivesProjectIntegrationWrite(id, notInstalled)).toBe(true);
+ }
+ });
+
+ test('a user-global-only editor never writes', () => {
+ for (const id of [
+ 'claude-desktop',
+ 'openclaw',
+ 'antigravity',
+ 'lm-studio',
+ 'hermes',
+ ] as const) {
+ expect(receivesProjectIntegrationWrite(id, installed)).toBe(false);
+ expect(receivesProjectIntegrationWrite(id, notInstalled)).toBe(false);
+ }
+ });
+
+ test('Copilot writes only once its user-global entry exists', () => {
+ // Copilot is skill-only at project scope (`.github/skills`) and
+ // `isProjectSkillPrerequisiteMet` refuses to write that skill until
+ // Copilot's user-global OpenKnowledge entry is there — the skill would sit
+ // on disk unloaded. Until then, a project setup writes nothing for it.
+ expect(EDITOR_PROJECT_CONFIG_PATH.copilot).toBeNull();
+ expect(USER_MCP_GATED_EDITOR_IDS.map(String)).toContain('copilot');
+ expect(receivesProjectIntegrationWrite('copilot', notInstalled)).toBe(false);
+ expect(receivesProjectIntegrationWrite('copilot', installed)).toBe(true);
+ });
+
+ test('every gated editor is skill-only — a project MCP config would make the gate moot', () => {
+ // The gate exists because the skill is the ONLY project artifact. An editor
+ // that also writes a project MCP config would short-circuit to true above,
+ // silently making its presence in the list decorative.
+ for (const id of USER_MCP_GATED_EDITOR_IDS) {
+ expect(EDITOR_PROJECT_CONFIG_PATH[id]).toBeNull();
+ expect(EDITOR_PROJECT_SKILL_ROOT[id]).not.toBeNull();
+ }
+ });
+});
diff --git a/packages/desktop/src/main/index.ts b/packages/desktop/src/main/index.ts
index 0b27a2621..6257f49c5 100644
--- a/packages/desktop/src/main/index.ts
+++ b/packages/desktop/src/main/index.ts
@@ -110,6 +110,7 @@ import {
initContent,
isProcessAlive,
normalizeFsPath,
+ ONBOARDING_BUNDLE_IDS,
prepareSingleFileOpen,
type ResolvedSkillHost,
RUNTIME_VERSION,
@@ -4455,6 +4456,32 @@ function buildReclaimUserSkillsOpts(): Parameters<typeof reclaimUserSkillsOnLaun
};
}
+/**
+ * Every destination a user-global bundle install writes to, tildified for
+ * display. Mirrors the reclaim's own destination set and BOTH its gates, so no
+ * surface can advertise a copy that will not be written:
+ *
+ * - `USER_SKILL_HOSTS`, not the project-shaped host list — that one drops
+ * Copilot and Pi by design, silently omitting `~/.copilot`, `~/.pi/agent`
+ * and `~/.gemini` from a list users read as complete.
+ * - `skillsRoot`, not `hostDir + '/skills'` — Pi's user root is
+ * `.pi/agent/skills`, which the naive shape renders as a nonexistent
+ * `~/.pi/skills`.
+ * - The `.agents` hub only when it already exists; the reclaim writes that
+ * copy but never creates the hub.
+ *
+ * Single source for the first-launch consent disclosure AND the Settings row,
+ * so the two cannot drift apart the way a hand-maintained second list did.
+ */
+function userGlobalSkillDestinations(home: string, name: string): string[] {
+ return [
+ ...(existsSync(join(home, AGENTS_HUB_DIR)) ? [`~/${AGENTS_SKILLS_ROOT}/${name}`] : []),
+ ...USER_SKILL_HOSTS.filter((h) => existsSync(join(home, h.hostDir))).map(
+ (h) => `~/${h.skillsRoot}/${name}`,
+ ),
+ ];
+}
+
function createMcpWiringOpts(opts: ArmMcpWiringOpts = {}) {
return {
isPackaged: app.isPackaged,
@@ -4493,37 +4520,20 @@ function createMcpWiringOpts(opts: ArmMcpWiringOpts = {}) {
return { ok: true as const };
},
},
- // Skills leg of the first-launch consent dialog: per-bundle rows for the
- // show payload + the confirm finalizer. `applyConsent` records every
- // bundle's decision, then reuses the launch reclaim (now decision-gated)
- // to install the enabled set and tear down any declined-but-present
- // bundle — one code path for install + removal.
+ // Skills leg of the first-launch consent dialog: the bundles onboarding
+ // offers + the confirm finalizer. `applyConsent` records each OFFERED
+ // bundle's decision, then reuses the launch reclaim (decision-gated) to
+ // install the enabled set and tear down any declined-but-present bundle —
+ // one code path for install + removal. Bundles outside the onboarding set
+ // are never touched here: no decision recorded, nothing installed, and an
+ // existing copy left alone.
skills: {
- computeDescriptors: () => {
- const home = osHomedir();
- // Resolve reach once — identical for every built-in. A ledger read can
- // throw (realpath containment), so degrade to zero hosts, which disables
- // the row's checkbox rather than dropping every skill row for the boot.
- let resolvedHosts: ResolvedSkillHost[];
- try {
- resolvedHosts = resolveBuiltinSkillHosts(home);
- } catch {
- resolvedHosts = [];
- }
- const hosts = resolvedHosts.map((h) => h.editor);
- return USER_GLOBAL_BUNDLE_IDS.map((id) => {
- const d = computeBuiltinSkillDisclosure(home, id);
- return {
- id,
- name: d.name,
- alreadyInstalled: d.installed,
- description: d.description,
- size: d.size,
- hosts,
- paths: d.paths,
- };
- });
- },
+ computeDescriptors: () =>
+ ONBOARDING_BUNDLE_IDS.map((id) => {
+ const home = osHomedir();
+ const name = BUNDLE_SKILL_NAME[id];
+ return { id, name, paths: userGlobalSkillDestinations(home, name) };
+ }),
applyConsent: async (enabledIds: readonly string[]) => {
const home = osHomedir();
// The consent dialog is the trust boundary: a failed decision write
@@ -4532,7 +4542,7 @@ function createMcpWiringOpts(opts: ArmMcpWiringOpts = {}) {
// to enabled, and still return ok — silently losing the user's
// decline. Surface {ok:false} so mcp-wiring defers the marker and the
// dialog re-fires for a retry (same as a failed PATH/editor write).
- for (const id of USER_GLOBAL_BUNDLE_IDS) {
+ for (const id of ONBOARDING_BUNDLE_IDS) {
try {
await writeBundleDecision(home, BUNDLE_SKILL_NAME[id], enabledIds.includes(id));
} catch (err) {
diff --git a/packages/desktop/src/main/mcp-wiring.test.ts b/packages/desktop/src/main/mcp-wiring.test.ts
index 25a7d3c20..98ffd1490 100644
--- a/packages/desktop/src/main/mcp-wiring.test.ts
+++ b/packages/desktop/src/main/mcp-wiring.test.ts
@@ -714,16 +714,19 @@ function buildWiringOpts(overrides: Partial<WiringOpts> = {}): WiringOpts {
}
describe('runMcpWiringOnFirstLaunch — skills consent leg', () => {
+ // Two bundles even though onboarding currently offers one: this module is
+ // generic over whatever main hands it, and a multi-bundle payload is what
+ // exercises the per-bundle grant/decline split.
const TWO_SKILLS = [
{
id: 'discovery',
name: 'open-knowledge-discovery',
- alreadyInstalled: false,
+ paths: ['~/.claude/skills/open-knowledge-discovery'],
},
{
id: 'write-skill',
name: 'open-knowledge-write-skill',
- alreadyInstalled: true,
+ paths: ['~/.claude/skills/open-knowledge-write-skill'],
},
];
@@ -765,6 +768,52 @@ describe('runMcpWiringOnFirstLaunch — skills consent leg', () => {
});
});
+ test('an omitted skills field records no decision and leaves installs standing', async () => {
+ // `undefined` is "no decision made", NOT "decline everything" — onboarding
+ // sends it when the user dismisses the setup offer, and an already-installed
+ // bundle must survive that. An empty array would tear one down.
+ const ipcMain = stubIpcMain();
+ const wc = fakeWebContents(11);
+ const fs = memoryFs();
+ const { cli } = buildFirstLaunchCli();
+ const skills = stubSkills(TWO_SKILLS);
+ const events: Array<Record<string, unknown>> = [];
+ const logger = {
+ info: () => {},
+ warn: () => {},
+ error: () => {},
+ event: (p: { event: string; [k: string]: unknown }) => events.push(p),
+ };
+ runMcpWiringOnFirstLaunch(
+ buildWiringOpts({ ipcMain, cli, fs, skills, logger, immediateDispatchTarget: wc }),
+ );
+
+ const confirm = ipcMain.handlers.get('ok:mcp-wiring:confirm');
+ const result = await confirm?.({ sender: { id: 11 } }, { editorIds: [] });
+ expect(result).toEqual({ ok: true });
+ expect(skills.consentCalls).toEqual([]);
+ expect(events.filter((e) => String(e.event).startsWith('mcp-wiring-skill-consent'))).toEqual(
+ [],
+ );
+ // The marker still lands — a decision about the OTHER legs was made.
+ expect(readMcpStatusMarker('/home/u', fs)).toMatchObject({ configured: true });
+ });
+
+ test('an explicit empty skills array still declines every offered bundle', async () => {
+ const ipcMain = stubIpcMain();
+ const wc = fakeWebContents(11);
+ const fs = memoryFs();
+ const { cli } = buildFirstLaunchCli();
+ const skills = stubSkills(TWO_SKILLS);
+ runMcpWiringOnFirstLaunch(
+ buildWiringOpts({ ipcMain, cli, fs, skills, immediateDispatchTarget: wc }),
+ );
+
+ const confirm = ipcMain.handlers.get('ok:mcp-wiring:confirm');
+ await confirm?.({ sender: { id: 11 } }, { editorIds: [], skills: [] });
+ expect(skills.consentCalls).toEqual([[]]);
+ });
+
test('a failed skills leg defers the marker so the dialog re-fires', async () => {
const ipcMain = stubIpcMain();
const wc = fakeWebContents(11);
@@ -1150,6 +1199,35 @@ describe('runMcpWiringOnFirstLaunch — PATH consent leg', () => {
expect(events.some((e) => e.event === 'mcp-wiring-path-descriptor-failed')).toBe(true);
});
+ test('a throwing skills descriptor degrades to no skill rows instead of killing the dialog', () => {
+ // Mirrors the PATH degradation above. `computeDescriptors` runs at ARMING
+ // time, before the IPC handlers register, so an unguarded throw would leave
+ // the user with no consent dialog and no way to reach one — a worse outcome
+ // than simply not offering the skill this boot.
+ const ipcMain = stubIpcMain();
+ const wc = fakeWebContents(11);
+ const events: Array<Record<string, unknown>> = [];
+ const skills = stubSkills([], {
+ computeDescriptors: () => {
+ throw new Error('bundle dir unreadable');
+ },
+ });
+ runMcpWiringOnFirstLaunch(
+ buildWiringOpts({
+ ipcMain,
+ skills,
+ immediateDispatchTarget: wc,
+ logger: { info() {}, warn() {}, error() {}, event: (e) => events.push(e) },
+ }),
+ );
+
+ expect(wc.sent).toHaveLength(1);
+ expect((wc.sent[0]?.payload as { globalSkills: unknown }).globalSkills).toEqual([]);
+ expect(events.some((e) => e.event === 'mcp-wiring-skill-descriptors-failed')).toBe(true);
+ // The dialog is still live: confirm is armed and answerable.
+ expect(ipcMain.handlers.has('ok:mcp-wiring:confirm')).toBe(true);
+ });
+
test('skip never touches the PATH surface', async () => {
const ipcMain = stubIpcMain();
const wc = fakeWebContents(11);
diff --git a/packages/desktop/src/main/mcp-wiring.ts b/packages/desktop/src/main/mcp-wiring.ts
index 51f6c63d8..16100b90a 100644
--- a/packages/desktop/src/main/mcp-wiring.ts
+++ b/packages/desktop/src/main/mcp-wiring.ts
@@ -354,9 +354,10 @@ export interface McpWiringPathInstallSurface {
* teardown pipeline in.
*/
export interface McpWiringSkillsSurface {
- /** One row per user-global bundle for the dialog. Read-only — must not
- * write. `alreadyInstalled` pre-checks the row (grandfathered install);
- * the disclosure fields drive the row's cost/reach/destination display. */
+ /** One row per bundle onboarding offers. Read-only — must not write.
+ * `paths` lists every destination the install writes to, so the dialog's
+ * disclosure comes from the installer's own gates rather than a parallel
+ * list. */
computeDescriptors(): McpWiringGlobalSkillDescriptor[];
/**
* Finalize the per-bundle decision. Records enabled for every id in
@@ -975,16 +976,18 @@ export function runMcpWiringOnFirstLaunch(opts: RunMcpWiringFirstLaunchOpts): Ru
}
// Skills leg — apply per-bundle consent when the dialog solicited a
- // decision (skill rows were offered). Runs after PATH and BEFORE the
- // marker write so a failed skills leg defers the marker like a failed
- // editor/PATH write: the dialog stays mounted for a same-boot retry and
- // re-fires next boot. `request.skills` is the checked subset; every offered
- // bundle not listed is recorded declined (and removed if present).
- if (skillDescriptors.length > 0) {
+ // decision (skill rows were offered) AND the user made one. Runs after PATH
+ // and BEFORE the marker write so a failed skills leg defers the marker like
+ // a failed editor/PATH write: the dialog stays mounted for a same-boot retry
+ // and re-fires next boot. An ARRAY `request.skills` is the checked subset —
+ // every offered bundle not listed is recorded declined (and removed if
+ // present). `undefined` means no decision was made, which is NOT the same as
+ // declining: the leg is skipped entirely, so nothing is recorded and an
+ // existing install is left standing. Onboarding relies on that distinction —
+ // dismissing the setup offer must never tear down a bundle already on disk.
+ if (skillDescriptors.length > 0 && Array.isArray(request?.skills)) {
const offeredIds = new Set(skillDescriptors.map((d) => d.id));
- const enabledIds = Array.isArray(request?.skills)
- ? request.skills.filter((id) => offeredIds.has(id))
- : [];
+ const enabledIds = request.skills.filter((id) => offeredIds.has(id));
let skillResult: { ok: true } | { ok: false; error: string };
try {
skillResult = await skills.applyConsent(enabledIds);
diff --git a/packages/desktop/src/main/path-install.test.ts b/packages/desktop/src/main/path-install.test.ts
index 7d527747b..a1219ad47 100644
--- a/packages/desktop/src/main/path-install.test.ts
+++ b/packages/desktop/src/main/path-install.test.ts
@@ -94,6 +94,7 @@ describe('ensureCliOnPath — consent gate (rc files are never written without a
const zshrc = readFileSync(join(h, '.zshrc'), 'utf8');
expect(zshrc).toContain('# >>> open-knowledge cli >>>');
expect(zshrc).toContain('Delete this whole block to opt out');
+ expect(existsSync(join(h, '.config', 'fish', 'conf.d', 'open-knowledge.fish'))).toBe(false);
const marker = readMarkerFile(h);
expect(marker.consent).toEqual({ status: 'granted', at: GRANTED.at });
const granted = events.find((e) => e.event === 'path-install-consent-granted');
@@ -379,6 +380,22 @@ describe('ensureCliOnPath', () => {
});
});
+ test('darwin bash creates .bash_profile from scratch', async () => {
+ const h = home();
+ const result = await ensureCliOnPath(
+ baseOpts(h, {
+ env: { HOME: h, SHELL: '/bin/bash' },
+ consentDecision: GRANTED,
+ }),
+ );
+ expect(result.status).toBe('installed');
+ expect(readFileSync(join(h, '.bash_profile'), 'utf8')).toContain(
+ '# >>> open-knowledge cli >>>',
+ );
+ expect(existsSync(join(h, '.zshrc'))).toBe(false);
+ expect(existsSync(join(h, '.config', 'fish', 'conf.d', 'open-knowledge.fish'))).toBe(false);
+ });
+
test('linux deb install: symlinks target the flat-layout wrapper and bash gets .bashrc', async () => {
const h = home();
const result = await ensureCliOnPath(
@@ -428,13 +445,40 @@ describe('ensureCliOnPath', () => {
if (result.status === 'failed-all') expect(result.error).toContain('EACCES');
});
- test('fish conf.d block uses fish syntax, not POSIX export', async () => {
+ test('fish login shell gets a fish-syntax conf.d block, not a POSIX block', async () => {
const h = home();
- await ensureCliOnPath(baseOpts(h, { consentDecision: GRANTED }));
+ await ensureCliOnPath(
+ baseOpts(h, {
+ env: { HOME: h, SHELL: '/opt/homebrew/bin/fish' },
+ consentDecision: GRANTED,
+ }),
+ );
const fish = readFileSync(join(h, '.config', 'fish', 'conf.d', 'open-knowledge.fish'), 'utf8');
expect(fish).toContain('# >>> open-knowledge cli >>>');
expect(fish).toContain('set -gx PATH');
expect(fish).not.toContain('export PATH');
+ expect(existsSync(join(h, '.zshrc'))).toBe(false);
+ });
+
+ test('a full pass retains a previously recorded Fish file for later cleanup', async () => {
+ const h = home();
+ const fishConf = join(h, '.config', 'fish', 'conf.d', 'open-knowledge.fish');
+ await ensureCliOnPath(
+ baseOpts(h, {
+ env: { HOME: h, SHELL: '/opt/homebrew/bin/fish' },
+ consentDecision: GRANTED,
+ }),
+ );
+
+ const newExe = '/Users/someone/Applications/OpenKnowledge.app/Contents/MacOS/OpenKnowledge';
+ await ensureCliOnPath(baseOpts(h, { executablePath: newExe }));
+
+ const marker = readMarkerFile(h);
+ expect(marker.rcFiles).toEqual([fishConf, join(h, '.zshrc')]);
+ expect(removePathShimFromRcFiles({ home: h, env: { SHELL: '/bin/zsh' } }).status).toBe(
+ 'removed',
+ );
+ expect(existsSync(fishConf)).toBe(false);
});
test('app update repoints canonical symlinks to the new bundle wrapper', async () => {
@@ -463,23 +507,105 @@ describe('computePathInstallDescriptor', () => {
});
expect(descriptor).toEqual({
shellDetected: true,
- rcFilesToTouch: ['~/.zshrc', '~/.config/fish/conf.d/open-knowledge.fish'],
+ rcFilesToTouch: ['~/.zshrc'],
alreadyInstalled: false,
});
});
- test('an existing .bash_profile joins the touch list; a non-zsh shell skips .zshrc creation', () => {
+ test('a fresh macOS Bash machine gets .bash_profile; an existing profile remains sufficient evidence', () => {
+ const freshHome = home();
+ expect(
+ computePathInstallDescriptor({
+ home: freshHome,
+ platform: 'darwin',
+ env: { SHELL: '/bin/bash' },
+ }).rcFilesToTouch,
+ ).toEqual(['~/.bash_profile']);
+
+ const configuredHome = home();
+ writeFileSync(join(configuredHome, '.bash_profile'), 'export BAR=1\n');
+ expect(
+ computePathInstallDescriptor({
+ home: configuredHome,
+ platform: 'darwin',
+ env: { SHELL: '/bin/sh' },
+ }).rcFilesToTouch,
+ ).toEqual(['~/.bash_profile']);
+ });
+
+ test('an unsupported shell with no recognized config has no PATH target', () => {
const h = home();
- writeFileSync(join(h, '.bash_profile'), 'export BAR=1\n');
+ expect(
+ computePathInstallDescriptor({
+ home: h,
+ platform: 'darwin',
+ env: { SHELL: '/bin/nu' },
+ }),
+ ).toEqual({ shellDetected: false, rcFilesToTouch: [], alreadyInstalled: false });
+ });
+
+ test('fish is targeted when it is the login shell or has independent usage evidence', () => {
+ const loginShellHome = home();
+ expect(
+ computePathInstallDescriptor({
+ home: loginShellHome,
+ platform: 'darwin',
+ env: { SHELL: '/opt/homebrew/bin/fish' },
+ }).rcFilesToTouch,
+ ).toEqual(['~/.config/fish/conf.d/open-knowledge.fish']);
+
+ const configuredHome = home();
+ mkdirSync(join(configuredHome, '.config', 'fish'), { recursive: true });
+ writeFileSync(join(configuredHome, '.config', 'fish', 'config.fish'), '# user config\n');
+ expect(
+ computePathInstallDescriptor({
+ home: configuredHome,
+ platform: 'darwin',
+ env: { SHELL: '/bin/zsh' },
+ }).rcFilesToTouch,
+ ).toEqual(['~/.zshrc', '~/.config/fish/conf.d/open-knowledge.fish']);
+
+ const usedHome = home();
+ mkdirSync(join(usedHome, '.config', 'fish'), { recursive: true });
+ writeFileSync(
+ join(usedHome, '.config', 'fish', 'fish_variables'),
+ '# This file contains fish universal variable definitions.\n',
+ );
+ expect(
+ computePathInstallDescriptor({
+ home: usedHome,
+ platform: 'darwin',
+ env: { SHELL: '/bin/zsh' },
+ }).rcFilesToTouch,
+ ).toEqual(['~/.zshrc', '~/.config/fish/conf.d/open-knowledge.fish']);
+ });
+
+ test('a legacy OK-created fish directory and conf file are not Fish-detection evidence', async () => {
+ const h = home();
+ const fishConf = join(h, '.config', 'fish', 'conf.d', 'open-knowledge.fish');
+ mkdirSync(dirname(fishConf), { recursive: true });
+ const legacyBlock =
+ '# >>> open-knowledge cli >>>\n# legacy OK-owned block\n# <<< open-knowledge cli <<<\n';
+ writeFileSync(fishConf, legacyBlock);
+
const descriptor = computePathInstallDescriptor({
home: h,
platform: 'darwin',
- env: { SHELL: '/bin/bash' },
+ env: { SHELL: '/bin/zsh' },
});
- expect(descriptor.rcFilesToTouch).toEqual([
- '~/.bash_profile',
- '~/.config/fish/conf.d/open-knowledge.fish',
- ]);
+ expect(descriptor.rcFilesToTouch).toEqual(['~/.zshrc']);
+ expect(descriptor.alreadyInstalled).toBe(false);
+
+ await ensureCliOnPath(baseOpts(h, { consentDecision: GRANTED }));
+ expect(readFileSync(fishConf, 'utf8')).toBe(legacyBlock);
+ expect(readMarkerFile(h).rcFiles as string[]).not.toContain(fishConf);
+
+ // Cleanup is deliberately broader than detection: Settings must still
+ // remove the self-identifying legacy file even when no marker owns it.
+ expect(removePathShimFromRcFiles({ home: h, env: { SHELL: '/bin/zsh' } }).status).toBe(
+ 'removed',
+ );
+ expect(existsSync(fishConf)).toBe(false);
});
test('a healthy managed block flips alreadyInstalled', async () => {
@@ -521,7 +647,6 @@ describe('computePathInstallDescriptor', () => {
await ensureCliOnPath(baseOpts(h, { consentDecision: GRANTED }));
// User strips the block → opt-out recorded on the next boot.
writeFileSync(join(h, '.zshrc'), 'export FOO=1\n');
- unlinkSync(join(h, '.config', 'fish', 'conf.d', 'open-knowledge.fish'));
await ensureCliOnPath(baseOpts(h));
const descriptor = computePathInstallDescriptor({
home: h,
@@ -565,6 +690,8 @@ describe('isPathShimInstalled / removePathShimFromRcFiles — the Settings → A
test('grant → installed reads true; remove strips every block, deletes the OK-owned fish conf, and records declined', async () => {
const h = home();
writeFileSync(join(h, '.zshrc'), 'export FOO=1\n');
+ mkdirSync(join(h, '.config', 'fish'), { recursive: true });
+ writeFileSync(join(h, '.config', 'fish', 'config.fish'), '# user fish config\n');
await ensureCliOnPath(baseOpts(h, { consentDecision: GRANTED }));
const env = { SHELL: '/bin/zsh' };
expect(isPathShimInstalled({ home: h, env })).toBe(true);
diff --git a/packages/desktop/src/main/path-install.ts b/packages/desktop/src/main/path-install.ts
index 0d507bf2b..655d82441 100644
--- a/packages/desktop/src/main/path-install.ts
+++ b/packages/desktop/src/main/path-install.ts
@@ -211,9 +211,24 @@ function rcTargets(
// the real host, where process.platform is the right answer.
platform: string = process.platform,
): Array<{ path: string; create: boolean; content: string }> {
+ const fishConfigDir = join(home, '.config', 'fish');
+ const fishDetected =
+ (shell?.endsWith('/fish') ?? false) ||
+ fs.existsSync(join(fishConfigDir, 'config.fish')) ||
+ // Fish creates fish_variables on first interactive use. Unlike the config
+ // directory and our conf.d file, older OpenKnowledge builds never created
+ // it, so it is independent evidence for users who launch Fish from a
+ // terminal profile while retaining zsh/bash as their login shell.
+ fs.existsSync(join(fishConfigDir, 'fish_variables'));
const base = [
{ path: join(home, '.zshrc'), create: shell?.endsWith('/zsh') ?? false, content: block() },
- { path: join(home, '.bash_profile'), create: false, content: block() },
+ {
+ path: join(home, '.bash_profile'),
+ // A fresh macOS Bash user needs `create: true`; otherwise the final
+ // filter only keeps entries for files that already exist on disk.
+ create: platform === 'darwin' && (shell?.endsWith('/bash') ?? false),
+ content: block(),
+ },
// Linux interactive non-login bash reads ~/.bashrc, not ~/.bash_profile
// (which most distros source only on login shells). Kept off macOS so
// the shipping darwin behavior — never touching a mac user's .bashrc —
@@ -227,11 +242,19 @@ function rcTargets(
},
]
: []),
- {
- path: join(home, '.config', 'fish', 'conf.d', 'open-knowledge.fish'),
- create: true,
- content: fishBlock(),
- },
+ // Fish gets the same current-shell-or-existing-user-config treatment as
+ // zsh/bash. Do NOT use the fish directory or this OK-owned conf.d file as
+ // evidence: older Desktop builds created both unconditionally, including
+ // on machines where Fish was never installed.
+ ...(fishDetected
+ ? [
+ {
+ path: join(fishConfigDir, 'conf.d', 'open-knowledge.fish'),
+ create: true,
+ content: fishBlock(),
+ },
+ ]
+ : []),
];
return base.filter((t) => t.create || fs.existsSync(t.path));
}
@@ -616,9 +639,14 @@ export async function ensureCliOnPath(opts: EnsureCliOnPathOpts): Promise<Ensure
if (canSkipRc && prior) {
rcFiles.push(...activePriorRcFiles);
} else if (rcConsented) {
+ // Keep no-longer-targeted prior files in the marker so Settings/uninstall
+ // can still clean up footprints from older targeting rules (notably the
+ // Fish file that older Desktop builds created for every user). Only the
+ // currently detected targets are written or refreshed.
+ rcFiles.push(...activePriorRcFiles);
for (const target of targets) {
if (upsertBlock(target.path, target.content, fs)) changedRcFiles.push(target.path);
- rcFiles.push(target.path);
+ if (!rcFiles.includes(target.path)) rcFiles.push(target.path);
}
} else {
// Declined or undecided: never write into the user's rc files. Keep
@@ -779,13 +807,18 @@ export function removePathShimFromRcFiles(opts: {
}): RemovePathShimResult {
const { home, fs = defaultFsOps, logger = DEFAULT_LOGGER } = opts;
const marker = readMarker(home, fs, logger);
+ const okOwnedFishConf = join(home, '.config', 'fish', 'conf.d', 'open-knowledge.fish');
const candidates = new Set<string>([
...rcTargets(home, (opts.env ?? process.env).SHELL, fs, opts.platform).map(
(target) => target.path,
),
...(marker?.rcFiles ?? []),
+ // Detection and cleanup intentionally differ: this self-identifying,
+ // OK-owned path is not evidence that Fish is installed, but uninstall
+ // must still find footprints from unconditional older installs even when
+ // their marker is missing or malformed.
+ okOwnedFishConf,
]);
- const okOwnedFishConf = join(home, '.config', 'fish', 'conf.d', 'open-knowledge.fish');
const strippedFiles: string[] = [];
try {
for (const path of candidates) {
diff --git a/packages/desktop/src/shared/ipc-channels.ts b/packages/desktop/src/shared/ipc-channels.ts
index 9e6544707..f876256de 100644
--- a/packages/desktop/src/shared/ipc-channels.ts
+++ b/packages/desktop/src/shared/ipc-channels.ts
@@ -504,29 +504,23 @@ export interface McpWiringPathInstallDescriptor {
export interface McpWiringConfirmRequest {
readonly editorIds: readonly McpWiringEditorId[];
readonly pathInstall?: boolean;
- /** Bundle ids the user left checked. Present ⇒ a skill decision was
- * solicited; every offered bundle not listed is recorded declined (and
- * removed if already installed). */
+ /** Bundle ids the user left checked. An ARRAY (even empty) ⇒ a skill decision
+ * was made; every offered bundle not listed is recorded declined (and removed
+ * if already installed). `undefined` ⇒ no decision was made — main skips the
+ * skills leg entirely, so nothing is written and nothing already installed is
+ * torn down. Onboarding sends `undefined` when the user declines the whole
+ * setup, because declining setup must never uninstall an existing bundle. */
readonly skills?: readonly string[];
}
-/** One user-global skill row in the first-launch consent dialog. Computed
- * read-only at arming time from `USER_GLOBAL_BUNDLE_IDS` + disk/marker state.
- * `alreadyInstalled: true` renders the row pre-checked as an existing install
- * the user can uncheck to remove. The disclosure fields mirror the Settings
- * skill status so both consent surfaces show the same what/cost/where; `hosts`
- * empty ⇒ nowhere to land and the row disables its checkbox. */
+/** One user-global skill bundle offered by the first-launch consent dialog.
+ * Computed read-only at arming time from `ONBOARDING_BUNDLE_IDS` + disk state.
+ * `paths` lists every destination the install writes to, computed from the
+ * installer's own iteration set + gates so a disclosure can never advertise a
+ * copy that will not be made. */
export interface McpWiringGlobalSkillDescriptor {
readonly id: string;
readonly name: string;
- readonly alreadyInstalled: boolean;
- /** The skill's own frontmatter description. */
- readonly description: string;
- /** Three-tier context cost; absent when the packaged bundle is unreadable. */
- readonly size?: SkillCostTiers;
- /** Resolved reach: static editor ids plus declared custom-root paths. */
- readonly hosts: readonly string[];
- /** Destination paths a confirmed install writes (or an uninstall removes). */
readonly paths: readonly string[];
}
diff --git a/packages/desktop/tests/smoke/mcp-wiring.e2e.ts b/packages/desktop/tests/smoke/mcp-wiring.e2e.ts
index 0812ac27c..1c3ac3136 100644
--- a/packages/desktop/tests/smoke/mcp-wiring.e2e.ts
+++ b/packages/desktop/tests/smoke/mcp-wiring.e2e.ts
@@ -330,7 +330,7 @@ test.describe('M6b first-launch MCP-wiring smoke (US-010)', () => {
}
});
- test('skip — writes configured:false marker and no editor configs', async ({
+ test('dismiss — writes configured:false marker and no editor configs', async ({
captureStderrFor,
}) => {
const tmpHome = createTmpHome('skip');
@@ -339,12 +339,16 @@ test.describe('M6b first-launch MCP-wiring smoke (US-010)', () => {
const app = await launchApp({ tmpHome });
captureStderrFor(app);
const window = await waitForConsentDialog(app);
- await window.getByTestId('mcp-consent-skip').click();
+ // Escape is the skip path now that the dialog has a single Continue
+ // button: unchecking says "don't set up", while dismissing says "I made
+ // no decision". Both must leave editor configs untouched, and only this
+ // one writes the skipped marker.
+ await window.keyboard.press('Escape');
await expect
.poll(() => readMarker(tmpHome), {
timeout: 15_000,
- message: 'skip marker not written within 15s of Skip click',
+ message: 'skip marker not written within 15s of dismissing the dialog',
})
.not.toBeNull();
diff --git a/packages/desktop/tests/smoke/qa-create-new-extended.e2e.ts b/packages/desktop/tests/smoke/qa-create-new-extended.e2e.ts
index 2036ab25b..3004e995e 100644
--- a/packages/desktop/tests/smoke/qa-create-new-extended.e2e.ts
+++ b/packages/desktop/tests/smoke/qa-create-new-extended.e2e.ts
@@ -38,8 +38,7 @@ import {
writeFileSync,
} from 'node:fs';
import { tmpdir } from 'node:os';
-import { join } from 'node:path';
-import { ALL_EDITOR_IDS } from '@inkeep/open-knowledge-core';
+import { delimiter, join } from 'node:path';
import type { ElectronApplication, Page } from '@playwright/test';
import { _electron as electron } from '@playwright/test';
import { typeProjectName } from './_helpers/create-new-dialog';
@@ -59,9 +58,9 @@ const TARGET = resolveDesktopTarget();
const DESKTOP_PRODUCT_NAME = '@inkeep/open-knowledge-desktop';
/**
- * Editor controls live inside the collapsed "Advanced settings" section (Radix
- * unmounts collapsed content) — expand it before driving or asserting on those
- * checkboxes. (Config sharing is top-level, not in Advanced.)
+ * Only the config-sharing posture lives inside the collapsed "Advanced settings"
+ * section now (Radix unmounts collapsed content) — expand it before driving or
+ * asserting on that control. The AI-tool decision is top-level.
*/
async function expandCreateAdvanced(page: Page): Promise<void> {
const trigger = page.locator('[data-testid="create-advanced-trigger"]');
@@ -69,6 +68,37 @@ async function expandCreateAdvanced(page: Page): Promise<void> {
await trigger.click();
}
+/**
+ * Make `bin` resolve as an installed CLI for editor-presence detection.
+ * `bin` must be the registry binary name (`TERMINAL_CLIS[<id>].bin` in
+ * core/src/handoff/terminal-launch.ts) — the editor id and its binary differ
+ * for some tools (cursor → `cursor-agent`).
+ * Detection is probe-based (login-shell `command -v` on POSIX, `where` on
+ * Windows) — a config directory under HOME is deliberately NOT a presence
+ * signal, so this seeds a real executable stub instead:
+ * - a stub under `<tmpHome>/bin` (`.cmd` flavor for `where` + PATHEXT),
+ * - login+interactive rc files prepending it to PATH (the POSIX probe spawns
+ * `$SHELL -l -i`, which rebuilds PATH from these; bash reads the profile
+ * files, zsh reads .zprofile/.zshrc),
+ * - launchApp additionally prepends `<tmpHome>/bin` to the app's PATH env,
+ * which is what the Windows `where` probe consults.
+ * Scheme-handler probes stay untouched — empty CI homes report none, which is
+ * the hermetic baseline these tests rely on.
+ */
+function seedCliOnPath(tmpHome: string, bin: string): void {
+ const binDir = join(tmpHome, 'bin');
+ mkdirSync(binDir, { recursive: true });
+ if (process.platform === 'win32') {
+ writeFileSync(join(binDir, `${bin}.cmd`), '@exit /b 0\r\n');
+ } else {
+ writeFileSync(join(binDir, bin), '#!/bin/sh\nexit 0\n', { mode: 0o755 });
+ const pathLine = 'export PATH="$HOME/bin:$PATH"\n';
+ for (const rc of ['.zprofile', '.zshrc', '.bash_profile', '.profile']) {
+ writeFileSync(join(tmpHome, rc), pathLine);
+ }
+ }
+}
+
function seedTmpHome(prefix: string, stateOverride?: Record<string, unknown>): string {
const tmpHome = realpathSync(mkdtempSync(join(tmpdir(), `ok-qa-${prefix}-`)));
const userDataDir = join(tmpHome, 'Library', 'Application Support', DESKTOP_PRODUCT_NAME);
@@ -105,6 +135,9 @@ async function launchApp(tmpHome: string, opts: LaunchOpts = {}): Promise<Electr
timeout: 30_000,
env: {
...process.env,
+ // `<tmpHome>/bin` first so seedCliOnPath stubs win resolution — the
+ // Windows `where` presence probe reads this PATH directly.
+ PATH: `${join(tmpHome, 'bin')}${delimiter}${process.env.PATH ?? ''}`,
...homeEnv(tmpHome),
OK_DESKTOP_E2E_SMOKE: '1',
...(opts.pickedParent !== undefined
@@ -173,8 +206,16 @@ test.describe('QA extended create-new-project', () => {
// editor customization → only chosen editors land on
// disk; aria roles; telemetry variant via flow_kind attribute.
- test('QA-005 editor customization writes only checked editors', async ({ captureStderrFor }) => {
+ test('QA-005 the AI-tool decision writes exactly the detected tools', async ({
+ captureStderrFor,
+ }) => {
const tmpHome = seedTmpHome('editors');
+ // Cursor only: the write set is now "every tool we detected", so what makes
+ // this test meaningful is that the OTHER editors' artifacts stay absent.
+ // Cursor's terminal CLI ships as `cursor-agent`, not `cursor` — must match
+ // `TERMINAL_CLIS.cursor.bin` (core/src/handoff/terminal-launch.ts) or the
+ // presence probe resolves nothing.
+ seedCliOnPath(tmpHome, 'cursor-agent');
const parent = join(tmpHome, 'projects');
mkdirSync(parent, { recursive: true });
const projectName = 'Customized';
@@ -205,18 +246,16 @@ test.describe('QA extended create-new-project', () => {
{ timeout: 5_000 },
);
- // The selection is seeded from detection, and the seeded tmp HOME has no
- // agent tool installed on it — so the baseline is nothing checked. Tick
- // cursor and only cursor. EDITOR IDs from cli (NOT 'claude-code').
- await expandCreateAdvanced(navigator);
- const cursorBox = navigator.locator('[data-testid="create-editor-cursor"]');
- await expect(cursorBox).not.toBeChecked();
- await expect(navigator.locator('[data-testid="create-editor-codex"]')).not.toBeChecked();
- await expect(navigator.locator('[data-testid="create-editor-claude"]')).not.toBeChecked();
- await cursorBox.click();
- await expect(cursorBox).toBeChecked();
- await expect(navigator.locator('[data-testid="create-editor-codex"]')).not.toBeChecked();
- await expect(navigator.locator('[data-testid="create-editor-claude"]')).not.toBeChecked();
+ // One pre-checked box covering the detected tools, top-level (no longer
+ // behind Advanced). Its label names the write set, so assert the label
+ // rather than per-editor checkboxes that no longer exist.
+ const connectBox = navigator.locator('[data-testid="create-editors-checkbox"]');
+ await expect(connectBox).toBeVisible({ timeout: 15_000 });
+ await expect(connectBox).toBeChecked();
+ const summary = navigator.locator('[data-testid="create-editors-summary"]');
+ await expect(summary).toContainText('Cursor');
+ await expect(summary).not.toContainText('Codex');
+ await expect(summary).not.toContainText('Claude');
const submit = navigator.locator('[data-testid="create-submit"]');
await expect(submit).toBeEnabled();
@@ -230,7 +269,7 @@ test.describe('QA extended create-new-project', () => {
.poll(() => existsSync(join(expected, '.ok', 'config.yml')), { timeout: 15_000 })
.toBe(true);
- // The ticked editor is wired; the unticked ones leave nothing behind.
+ // The detected editor is wired; undetected ones leave nothing behind.
await expect
.poll(() => existsSync(join(expected, '.cursor', 'mcp.json')), { timeout: 15_000 })
.toBe(true);
@@ -239,10 +278,11 @@ test.describe('QA extended create-new-project', () => {
expect(existsSync(join(expected, '.mcp.json'))).toBe(false);
});
- // dialog UX: name input is focused on open, Location is hydrated,
- // every editor row is visible and tickable under Advanced settings.
+ // dialog UX: name input is focused on open, Location is hydrated, and the
+ // AI-tool decision is a single top-level checkbox with a live status region.
test('QA-010 dialog UX — focus, location, checkboxes, ARIA', async ({ captureStderrFor }) => {
const tmpHome = seedTmpHome('uxshape');
+ seedCliOnPath(tmpHome, 'codex');
const parent = join(tmpHome, 'projects');
mkdirSync(parent, { recursive: true });
const projectName = 'Live Preview';
@@ -271,19 +311,30 @@ test.describe('QA extended create-new-project', () => {
const ariaLive = await caption.getAttribute('aria-live');
expect(ariaLive).toBe('polite');
- // Every known editor renders a row under Advanced settings. None starts
- // checked: the selection is seeded from detection and nothing is installed
- // under the seeded tmp HOME.
- await expandCreateAdvanced(navigator);
- for (const id of ALL_EDITOR_IDS) {
- const checkbox = navigator.locator(`[data-testid="create-editor-${id}"]`);
- await expect(checkbox).toBeVisible();
- await expect(checkbox).not.toBeChecked();
- }
- // The rows are interactive, not merely rendered.
- const codexBox = navigator.locator('[data-testid="create-editor-codex"]');
- await codexBox.click();
- await expect(codexBox).toBeChecked();
+ // The AI-tool decision is one top-level, pre-checked box naming the tools
+ // it covers — no longer a row per editor behind Advanced settings.
+ const connectBox = navigator.locator('[data-testid="create-editors-checkbox"]');
+ await expect(connectBox).toBeVisible({ timeout: 15_000 });
+ await expect(connectBox).toBeChecked();
+ await expect(navigator.locator('[data-testid="create-editors-summary"]')).toContainText(
+ 'Codex',
+ );
+
+ // The status region is a live region that persists across probe states, so
+ // "checking…" → settled actually announces.
+ const status = navigator.locator('[data-testid="create-editors-status"]');
+ expect(await status.getAttribute('aria-live')).toBe('polite');
+ await expect(status).toHaveAttribute('data-status', 'ready');
+
+ // The disclosure names the exact project files, and the box is interactive.
+ await navigator.locator('[data-testid="create-editors-details-toggle"]').click();
+ await expect(navigator.locator('[data-testid="create-editors-details"]')).toContainText(
+ '.codex/config.toml',
+ );
+ await connectBox.click();
+ await expect(connectBox).not.toBeChecked();
+ await connectBox.click();
+ await expect(connectBox).toBeChecked();
// Type the name + Browse → caption shows the resolved target.
await typeProjectName(navigator, projectName);
@@ -393,8 +444,9 @@ test.describe('QA extended create-new-project', () => {
rmSync(join(userDataDir, 'bug-report-dirty-shutdown.json'), { force: true });
// Second launch: relaunch. Location prefills from lastUsedProjectParent;
- // the Name input resets to empty; the editor selection re-seeds from
- // detection (still empty under the seeded tmp HOME).
+ // the Name input resets to empty; the AI-tool decision re-probes detection
+ // (which finds nothing under the seeded tmp HOME) and the checkbox itself
+ // resets to its pre-checked default.
const app2 = await launchApp(tmpHome);
captureStderrFor(app2);
const navigator2 = await findWindowByMode(app2, 'navigator', 30_000);
@@ -411,13 +463,18 @@ test.describe('QA extended create-new-project', () => {
persistedParent,
{ timeout: 15_000 },
);
+ // Nothing is installed under this HOME, so the probe settles empty and the
+ // row says so instead of offering a checkbox that would write nothing.
+ await expect(navigator2.locator('[data-testid="create-editors-status"]')).toHaveAttribute(
+ 'data-status',
+ 'none',
+ { timeout: 15_000 },
+ );
+ await expect(navigator2.locator('[data-testid="create-editors-checkbox"]')).toHaveCount(0);
+ // Advanced still collapses back to hidden on a fresh open.
+ await expect(navigator2.locator('[data-testid="create-sharing"]')).toHaveCount(0);
await expandCreateAdvanced(navigator2);
- await expect(navigator2.locator('[data-testid="create-editor-claude"]')).not.toBeChecked();
- await expect(
- navigator2.locator('[data-testid="create-editor-claude-desktop"]'),
- ).not.toBeChecked();
- await expect(navigator2.locator('[data-testid="create-editor-cursor"]')).not.toBeChecked();
- await expect(navigator2.locator('[data-testid="create-editor-codex"]')).not.toBeChecked();
+ await expect(navigator2.locator('[data-testid="create-sharing"]')).toBeVisible();
});
// Create stays enabled even before the user types a name — a disabled
diff --git a/packages/server/src/index.ts b/packages/server/src/index.ts
index c658fee03..4d71f800e 100644
--- a/packages/server/src/index.ts
+++ b/packages/server/src/index.ts
@@ -529,6 +529,7 @@ export {
BUNDLE_SCOPE,
BUNDLE_SKILL_NAME,
bundleSkillMdPath,
+ ONBOARDING_BUNDLE_IDS,
USER_GLOBAL_BUNDLE_IDS,
} from './skill-bundles.ts';
export {
diff --git a/packages/server/src/skill-bundles.test.ts b/packages/server/src/skill-bundles.test.ts
index 9d4adfac0..f3bd3849b 100644
--- a/packages/server/src/skill-bundles.test.ts
+++ b/packages/server/src/skill-bundles.test.ts
@@ -5,8 +5,10 @@ import { extractSkillRefs, RESERVED_PROJECT_SKILL_NAME } from '@inkeep/open-know
import { describe, expect, test } from 'vitest';
import {
BUNDLE_IDS,
+ BUNDLE_SCOPE,
BUNDLE_SKILL_NAME,
bundleSkillMdPath,
+ ONBOARDING_BUNDLE_IDS,
USER_GLOBAL_BUNDLE_IDS,
} from './skill-bundles.ts';
@@ -58,6 +60,27 @@ describe('skill-bundles (single source of truth)', () => {
}
});
+ test('onboarding offers a non-empty subset of the user-global bundles', () => {
+ // The constant is a literal (a `.filter()` would widen to `BundleId[]` and
+ // let a bundle rename silently empty it), so the two properties a filter
+ // used to give for free are asserted here: every onboarding bundle is
+ // actually user-global, and the set is never empty — an empty one would
+ // mean first launch quietly stopped offering any skill.
+ expect(ONBOARDING_BUNDLE_IDS.length).toBeGreaterThan(0);
+ for (const id of ONBOARDING_BUNDLE_IDS) {
+ expect(USER_GLOBAL_BUNDLE_IDS).toContain(id);
+ expect(BUNDLE_SCOPE[id]).toBe('user');
+ }
+ });
+
+ test('write-skill is deliberately NOT an onboarding bundle', () => {
+ // Onboarding covers what decides whether a tool works at all; authoring
+ // helpers wait for Settings or an explicit `ok init`. Recorded here because
+ // the absence is the decision, and absences do not fail loudly on their own.
+ expect(USER_GLOBAL_BUNDLE_IDS).toContain('write-skill');
+ expect(ONBOARDING_BUNDLE_IDS as readonly string[]).not.toContain('write-skill');
+ });
+
test('write-skill description is within the skill contract (≤1024, no XML tags)', () => {
const raw = readFileSync(join(REPO_ROOT, bundleSkillMdPath('write-skill')), 'utf-8');
const desc = /description:\s*"([\s\S]*?)"\n/.exec(raw)?.[1] ?? '';
diff --git a/packages/server/src/skill-bundles.ts b/packages/server/src/skill-bundles.ts
index dafef3ba4..da38a2a69 100644
--- a/packages/server/src/skill-bundles.ts
+++ b/packages/server/src/skill-bundles.ts
@@ -69,6 +69,26 @@ export const BUNDLE_SCOPE: Record<BundleId, 'user' | 'project'> = {
*/
export const USER_GLOBAL_BUNDLE_IDS = BUNDLE_IDS.filter((id) => BUNDLE_SCOPE[id] === 'user');
+/**
+ * The user-global bundles first-launch onboarding sets up as part of its single
+ * "connect my AI tools" decision. `discovery` only: it is what lets an agent
+ * recognize an OpenKnowledge project at all, so it belongs with the MCP wiring.
+ * `write-skill` is an authoring convenience with no bearing on whether the tools
+ * work, and onboarding neither installs nor records a decision for it — it stays
+ * available from Settings and from `ok init`.
+ *
+ * A bundle absent here is left with NO recorded decision, which `resolveBundleEnabled`
+ * reads as "uninstalled on a fresh machine, untouched where it already exists" —
+ * so dropping one from onboarding never uninstalls an existing copy.
+ *
+ * Spelled as a literal rather than filtered out of `USER_GLOBAL_BUNDLE_IDS`: a
+ * filter widens to `BundleId[]`, so renaming the `discovery` bundle would
+ * silently yield an EMPTY onboarding set — first launch would quietly stop
+ * offering the skill. `satisfies` turns that rename into a type error instead.
+ * The subset relationship with `USER_GLOBAL_BUNDLE_IDS` is asserted in tests.
+ */
+export const ONBOARDING_BUNDLE_IDS = ['discovery'] as const satisfies readonly BundleId[];
+
/**
* Repo-relative path of a bundle's `SKILL.md` (the file whose
* `metadata.version` the release sync bumps). Derived from the id, which equals