Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image input.
C++
/**
* The agent preset the app's chats run under: the Harness's `standard` preset (pinned 57190202,
* packages/bundle/web-app/presets/standard.patch.yml) with the differences the app needs, and nothing else.
*
* A patch replaces the whole preset row, so the list is restated here. `tests/agent-tools.test.mjs` compares this list
* with the pinned Harness's own file (when STRATA_HARNESS_DIR is set), so a Harness update that adds or renames a
* plugin fails a test instead of silently dropping a tool.
*
* Differences from `standard`, each with its reason:
* - `tool-schedule` removed: the app has one scheduler (server/tasks.mjs, the Tasks screen); a second set of
* reminder tools would run behind its back.
* - `web_search` removed (tool-web with `search:false`): every search backend the Harness has needs a service key
* the app does not have, so the tool could only fail. `web_fetch` stays. (`searchProvider` can switch it back on.)
* - `tool-ask-user` removed: the app has no place to show the agent's own questions, so the tool could only fail;
* the agent asks in plain text instead.
* - the `planning` group (Harness plan mode and `exit_plan_mode`) removed: the app's Plan mode is the permission
* preset (nothing can change); the Harness's plan mode needs an interactive review the app does not have.
* - `tool-bash` / `tool-pwsh`: exactly one, by operating system (the Harness chooses the same at load).
* - `tool-plugin-manager` stays disabled, as in `standard`.
* - The goal, skill, workflow and fork tools are kept (owner decision): the first message is slower, and the short plain system prompt below keeps it as small as it can be.
*/
import { withoutDisabled } from './tools-catalog.mjs';
/** Plugin ids of the pinned Harness `standard` preset, in its order (the drift check). */
export const STANDARD_PLUGIN_IDS = [
'persona', 'agent-instructions', 'time-context', 'tool-bash', 'tool-pwsh', 'tool-fs', 'tool-fs-search', 'tool-jobs',
'tool-schedule', 'skill-filesystem', 'tool-skill', 'command-goal', 'tool-goal', 'planning', 'compaction', 'delegation',
'tool-ask-user', 'tool-todo', 'tool-web', 'present', 'tool-plugin-manager',
];
/** What we leave out of it. */
export const REMOVED_PLUGIN_IDS = ['tool-schedule', 'tool-ask-user', 'planning'];
/** Plugins removed from inside the `delegation` group (the group stays: helper agents are a feature). */
export const REMOVED_DELEGATION_IDS = [];
/**
* The whole system prompt of an app chat. The Harness's own text is about 4.8 KB (about 1,200 tokens) of guidance for working on
* the Harness itself (its checkout, its web page, its plugins): a local model pays for it on every first message and the person
* can see its words in the chat ("DSH", "sandbox"). This one says what a chat needs and nothing else. The tools state their own rules.
*/
export const AGENT_PROMPT = [
'You are a helpful assistant that can also work on this computer through your tools. You are running as {{model}}.',
'Answer questions directly and briefly; use a tool only when the task needs one. The person reads your replies in a chat: use plain words, not tool names or technical jargon.',
'Look before you change: read a file before you edit or overwrite it, and prefer a small edit to a rewrite. Search with the glob and grep tools and read files with the read tool; use the shell for running programs.',
'A command that starts a program, folder or page only tells you that it started. Never say a window opened, a page rendered or that you checked how it looks unless a tool confirmed it. If a command fails, say what went wrong in one sentence and what you will try next.',
'To show the person an image, SVG, HTML page or PDF from their folders, use the present tool: the app shows it next to the chat. A browser cannot open file:// addresses. If you do start a program, start it so the command returns at once and never wait for it to close: Start-Process on Windows, open on macOS, xdg-open on Linux.',
'When the person asks you to ask a subagent, a helper or another model ("ask a subagent", "ask model X"), use the subagent tool and name the model in its model field. Never look for a model with shell commands, and never wait for a helper with sleep: end your turn; its answer arrives as a new message.',
'Tokens prefixed with @ are files or folders the person pointed at. A failed or refused action changed nothing; say so, and do not claim work you did not do.',
].join('\n\n');
/**
* @param {{platform?: string, search?: boolean}} [options]
* @returns {object} the `config` of the `preset-standard` row
*/
export function standardPreset({ platform = process.platform, search = false, off = new Set(), skillsDir = null, compaction = [] } = {}) {
const windows = platform === 'win32';
const plugins = [
{ id: 'persona', name: '@deepseek-ai/dsh-persona', config: { prefix: AGENT_PROMPT, complete: true, includeRuntimeContext: false } },
{ id: 'agent-instructions', name: '@deepseek-ai/dsh-agent-instructions', config: { maxBytes: 65536 } },
{ id: 'time-context', name: '@deepseek-ai/dsh-time-context' },
windows
? { id: 'tool-pwsh', name: '@deepseek-ai/dsh-tool-pwsh' }
: { id: 'tool-bash', name: '@deepseek-ai/dsh-tool-bash' },
{ id: 'tool-fs', name: '@deepseek-ai/dsh-tool-fs' },
{ id: 'tool-fs-search', name: '@deepseek-ai/dsh-tool-fs-search', config: { sampleOverCapGlobResults: false } },
{ id: 'tool-jobs', name: '@deepseek-ai/dsh-tool-jobs' },
// The skills of the Skills page, written out as folders the Harness scans (and watches: a skill saved in the app is there at once).
{ id: 'skill-filesystem', name: '@deepseek-ai/dsh-skill-filesystem', ...(skillsDir ? { config: { customSkillDirs: [skillsDir] } } : {}) },
{ id: 'tool-skill', name: '@deepseek-ai/dsh-tool-skill' },
{ id: 'command-goal', name: '@deepseek-ai/dsh-command-goal' },
{ id: 'tool-goal', name: '@deepseek-ai/dsh-tool-goal' },
{
id: 'compaction', name: 'cordis:group', group: true, isolate: { compaction: true, toolResultPruner: true },
config: [
{ id: 'compaction-basic', name: '@deepseek-ai/dsh-compaction-basic', ...(compaction.length ? { config: { modelPolicies: compaction } } : {}) },
{ id: 'command-compact', name: '@deepseek-ai/dsh-command-compact' },
{ id: 'tool-result-pruner', name: '@deepseek-ai/dsh-compaction-tool-result-pruner', config: { thresholdChars: 8192, headChars: 4096, tailChars: 1024 } },
],
},
{
id: 'delegation', name: 'cordis:group', group: true, isolate: { workflowEngine: true },
config: [
{ id: 'tool-subagent-control', name: '@deepseek-ai/dsh-tool-subagent-control' },
{ id: 'tool-subagent-list-agents', name: '@deepseek-ai/dsh-tool-subagent-control/list-agents' },
{ id: 'tool-subagent', name: '@deepseek-ai/dsh-tool-subagent', config: { provider: 'spawn', toolName: 'subagent', modelSelectionSettings: true } },
{ id: 'tool-subagent-fork', name: '@deepseek-ai/dsh-tool-subagent', config: { provider: 'fork', toolName: 'subagent_fork' } },
{ id: 'workflow-ptc', name: '@deepseek-ai/dsh-workflow-ptc', config: { provider: 'spawn' } },
{ id: 'tool-workflow', name: '@deepseek-ai/dsh-tool-workflow' },
],
},
{ id: 'tool-todo', name: '@deepseek-ai/dsh-tool-todo', config: { allowParallelInProgress: true } },
{ id: 'tool-web', name: '@deepseek-ai/dsh-tool-web', config: { fetch: !off.has('fetch'), search: search && !off.has('search'), searchTimeoutMs: 60000 } },
{ id: 'present', name: '@deepseek-ai/dsh-tool-present' },
{ id: 'tool-plugin-manager', name: '@deepseek-ai/dsh-plugin-manager/tools', disabled: true },
];
// The groups the person switched off in Settings -> Tools; a web plugin with both halves off is not loaded at all.
const kept = withoutDisabled(plugins, off).filter(p => p.id !== 'tool-web' || p.config.fetch || p.config.search);
return { id: 'standard', order: 1, plugins: kept };
}