Skip to content

armgit / Pulse

Public Imported from GitHub
0

Real-time monitoring dashboard for Proxmox VE, PBS, Docker, Kubernetes, TrueNAS and vSphere. Self-hosted, with alerts for failed backups, full disks, unhealthy containers and offline nodes.

Make weekly Patrol summary help usable and evidence-aware

pulse-triage[bot]04493c7490dc
diff --git a/docs/AI.md b/docs/AI.md
index 5ee09e222f..bce42d5bd5 100644
--- a/docs/AI.md
+++ b/docs/AI.md
@@ -494,6 +494,122 @@ When control level is **Controlled**, write actions pause for approval:
---
+## Weekly summary and email
+
+Pulse's **This week** summary adds up retained Patrol activity over the last
+seven days. Open **Patrol → Activity** to read it; use **Settings → Reporting**
+to schedule the same summary by email. The card, weekly email and **Report type**
+selector are included in **v6.5.0**. An older build can lack these controls.
+
+This is a summary of recorded activity, not a complete audit or proof that every
+resource is healthy. Use individual findings, run history and **Actions** for
+the result of a particular investigation or fix.
+
+### Read the summary
+
+The card normally says **Last 7 days**: a rolling window, not a calendar week.
+If it says **Since … (older runs are no longer kept)**, its run history covers
+only part of that window. Refreshing the card reads existing records; it does
+not run Patrol or recover missing history.
+
+| Line | What it means and what to check |
+| --- | --- |
+| **Patrol runs** | Retained runs, checks and failures in the window. The resource count is the largest number checked in one run, not a count of distinct resources across the week. Read run history for failed or missing checks. |
+| **New issues** | New-finding totals recorded by those runs. The open severity counts concern findings detected within the window; older open findings are not included. Check the full findings list before concluding that nothing needs attention. |
+| **Issues resolved** | Recorded automatic resolutions plus manual resolutions still retained. Dismissed and muted findings are described separately. A cleared finding does not establish that an action ran or was independently verified. |
+| **Investigated** | Recorded investigation outcomes on retained findings. “Fix run, result unknown” is not a verified fix; open the finding for its result and remaining work. |
+| **Fixes run** | Actions recorded as successfully executed. “Verified afterwards” is a separate count; review failed actions and waiting approvals in **Actions**. Do not run an action again merely to increase the count. |
+| **Estimated spend** | Recorded Patrol model calls and their estimated USD cost, not an invoice or all Assistant usage. If some calls have no known price, the displayed amount is incomplete, not a zero-cost result. Check [Cost Tracking](AI.md#cost-tracking). |
+
+The displayed mode is Patrol's **current** mode, not a history of mode changes.
+Watch-only mode can explain why investigations or fixes were not performed,
+but changing mode is not a repair for an empty summary.
+
+### Schedule the weekly email
+
+Scheduling requires the **advanced reporting** entitlement and an account
+allowed to manage report schedules. Reading the Patrol card does not itself
+grant scheduling access. See [Plans and entitlements](PULSE_PRO.md).
+
+1. Check that email notifications are **enabled and configured** under
+ **Alerts → Notifications**. Listing recipients on a schedule does not
+ configure a mail provider or enable email.
+2. Open **Settings → Reporting**, choose **Create schedule**, then set
+ **Report type** to **Patrol weekly summary**.
+3. Enter a schedule name, weekday, time and an IANA **Timezone**, such as
+ `Europe/London`. The summary is weekly, covers the last seven days, and is
+ sent as HTML and plain text email. It is not a PDF/CSV attachment, a
+ resource/tag-filtered report, or a saved disk copy.
+4. Set **Email recipients** deliberately. Blank uses the existing email
+ notification recipients; if none are configured, delivery can fall back
+ to the sender address. Use explicit authorised recipients to avoid that
+ ambiguity. The email can contain resource names, findings and costs.
+5. Save with **Enabled** selected for automatic runs. **Run now** sends a real
+ email; it is not a preview or a test of future scheduled delivery. Use it
+ only when those recipients should receive the summary.
+
+The Report type choice appears when the session reports AI as enabled. An
+existing summary keeps its type when edited after AI is switched off. If the
+choice is missing, check the installed version, current workspace, AI settings
+and reporting access; do not change Patrol's autonomy mode, re-enrol agents or
+create an API token just to reveal it.
+
+### Check a missing email or empty summary
+
+- **No recorded runs:** check Patrol's enabled state, configured provider,
+ ordinary run history and errors. “Patrol has not run” can also reflect an
+ unavailable Patrol service, not proof that the environment had no problems.
+ Do not force a run, restart Pulse or change retention just to fill the card.
+- **Failed schedule:** read **Last run** and the saved error in Reporting. On
+ a narrow screen, use the same signed-in page at a wider viewport to see
+ that column. A missing Patrol service and missing email configuration are
+ different failures. Preserve the original error; do not repeatedly use
+ Run now or send synthetic alerts to diagnose them.
+- **Run marked OK but no email received:** that status means Pulse's sender
+ returned success, not that the message reached the recipient's inbox.
+ Check the selected recipients and the mail provider's delivery or rejection
+ record, then the mailbox's filtering. A successful alert Test does not prove
+ that this schedule ran or used the intended recipients.
+- **Email links missing:** links depend on the configured Pulse public URL.
+ Open Patrol in your existing authenticated browser; do not make Pulse
+ publicly accessible or disable TLS to make an email link work.
+
+Keep full mail errors, recipient addresses, credentials and exports private.
+Share only a locally reviewed, redacted error and the relevant schedule time
+and outcome; see [Getting Help](TROUBLESHOOTING.md#-getting-help).
+
+### Honest limits of the data
+
+- Run history retains at most **100 runs**. Event-triggered activity can
+ exhaust it before seven days. The partial-window notice concerns run
+ retention only; its absence does not certify all other lines as complete.
+- Resolved findings can be purged after **24 hours**, and dismissed findings
+ after **30 days**. Manual resolutions, dismissals and investigation outcomes
+ on purged findings can disappear from these totals. Automatic resolution
+ totals come from retained runs, so the lines need not reconcile exactly.
+- Actions have their own retained audit records, but the digest reads a bounded
+ set and can show an empty action line when that source is unavailable.
+ Check **Actions** itself; zero in this summary is not proof that no action
+ took place. Pending approvals can include actions created before the window.
+- Model spend depends on recorded usage and known prices. This summary cannot
+ establish that caching, a budget limit or a provider's bill is correct.
+
+### API
+
+For an existing authenticated integration, `GET /api/ai/patrol/digest` requires
+`ai:execute` scope. Optional `days` is an integer from **1 to 30**, default
+**7**; invalid values return 400. Scheduled email always uses seven days.
+The response groups `window`, `mode`, `runs`, `findings`, `investigations`,
+`actions`, `alerts` and `spend`. `window.history_complete` and
+`window.history_since` describe the run-retention limitation above, not a
+whole-report completeness guarantee.
+
+A zero response can be returned when the Patrol service is unavailable.
+Authentication, scope or transport errors are not a zero result. Prefer the
+signed-in card for a one-off check; never extract a session cookie or put an
+API token in a command, URL or public report. See [API authentication](API.md#-authentication)
+for the private header-file approach when an integration needs it.
+
## Configuration
Configure providers in the UI: **Settings → Pulse Intelligence → Provider & Models**
diff --git a/docs/PATROL_WEEKLY_DIGEST.md b/docs/PATROL_WEEKLY_DIGEST.md
index bcca2943e4..939236cea6 100644
--- a/docs/PATROL_WEEKLY_DIGEST.md
+++ b/docs/PATROL_WEEKLY_DIGEST.md
@@ -1,105 +1,6 @@
# Patrol Weekly Digest
-Status: building — endpoint and in-app "This week" card on `main` (PRs #1856
-and #1860, 2026-09-02); the weekly email is the `patrol_digest` report
-schedule kind, and Settings > Reporting creates it through the "Report type"
-choice on the schedule form (2026-10-02). Demand ledger:
-`pulse-pro/FEATURE_REQUESTS.md`, "Patrol weekly digest (what Patrol did for
-you)", a named bet under the Patrol operations loop.
-
-## The job, in the customer's words
-
-"Show me what I am paying for." A Pro customer turns Patrol on, it runs about
-five times a day, and nothing in Pulse ever adds that up. They see individual
-findings when they open the page and individual emails when a finding fires,
-but never the week: how often Patrol looked, what it caught, what it fixed,
-what it cost. The 2026-09-01 assessment found 33 of 175 Pro subscriptions past
-due and 80 percent of paying installs never seeing the paid loop fire. A
-customer who cannot see the work stops paying for it.
-
-The least-expert plausible reader is a homelab operator who set Patrol up once
-and opens Pulse when an alert email arrives. They do not know what an
-"investigation", a "verdict", or an "evidence class" is. The digest must read
-as a plain weekly account, and every number must lead to something they can
-click or a decision they can make.
-
-## What the digest reports
-
-Everything below is computed from data Pulse already records. No new
-telemetry fields; the readout task grades this bet from support and issue
-mentions.
-
-| Line | Source | What the reader does with it |
-| --- | --- | --- |
-| Runs: how many times Patrol checked, how many checks in total, how many resources it covered, when it last ran | `PatrolRunRecord` history (`internal/ai/patrol.go`, capped at 100 runs) | Confirms Patrol is alive. If runs are zero or failed, the card says so and points at Patrol setup. |
-| Mode: monitor, approval, assisted, or full | `Service.GetEffectivePatrolAutonomyLevel` | Explains why actions were or were not taken. Monitor mode with fixable findings is the upgrade prompt, stated plainly. |
-| New issues this week, and how many are still open by severity | run `new_findings` totals plus the findings store for still-tracked findings detected in the window | Open critical or warning issues are the thing to look at; the count links to the findings list. |
-| Issues resolved and how many Patrol resolved on its own | run `resolved_findings` totals plus manual resolutions still in the findings store | This is the "it did something" line. |
-| Issues you dismissed or muted | finding lifecycle events (`dismissed`, `suppressed`) in the window | Reminds the reader what they told Patrol to ignore. |
-| Investigations and their outcomes | finding lifecycle `investigation_outcome` events in the window | Shows the paid loop firing; outcomes that need attention are named. |
-| Actions proposed, approved, executed, verified, failed, and still waiting | canonical action audit records with `origin.surface = patrol` (`internal/unifiedresources`) | "Waiting" links straight to Actions and approvals. |
-| Alerts Patrol looked into | alert-triggered runs (`alert_fired`, `alert_flapping`) and their distinct alert identifiers | Shows Patrol responding to the reader's own alerts, not just its schedule. |
-| Estimated model spend, tokens, and calls | `internal/ai/cost` usage events with `use_case = patrol` | Answers "what does this cost me" without opening the AI cost dashboard. Marked as an estimate; unknown pricing is stated, never zeroed. |
-
-Deliberately left out: per-run tool call traces, evidence classes, verdict
-strings, model names, and anything the reader cannot act on. Those stay in run
-history and the Actions audit, one click away.
-
-## Honest limits of the data
-
-- Run history keeps the last 100 runs. At five or six runs a day a week fits,
- but a busy install with event-triggered runs can overflow. The payload
- carries `window.history_complete` and `window.history_since`; the card says
- "since <date>" instead of "this week" when the window is cut short.
-- Resolved findings are purged from the findings store 24 hours after
- resolution, and dismissed findings after 30 days. Resolution totals therefore
- come from run records, and severity is only reported for findings still
- tracked. Investigation outcomes are read from lifecycle events on tracked
- findings, so a finding that was investigated, fixed, and purged more than a
- day ago no longer contributes. Executed and verified actions do not have this
- gap: action audit records are durable.
-- Spend is `cost.EstimateUSD` over recorded usage events. When the model has
- no known price the digest says pricing is unknown rather than reporting a
- smaller number.
-
-## Surface
-
-**In-app card first.** A "This week" card at the top of the Patrol page's
-Activity tab, above Verified outcomes
-(`frontend-modern/src/features/patrol/PatrolWeeklyDigestCard.tsx`, backed by
-`GET /api/ai/patrol/digest?days=7`). It costs nothing to deliver, every paying
-install can see it, and the Activity tab is already where "what happened"
-questions are answered. The Inbox stays a decision surface and does not gain a
-summary card.
-
-**Weekly email second.** The past-due population is the population that has
-stopped opening Pulse, so the email is the slice that reaches them. It is a
-report schedule kind, `patrol_digest`, on the report schedule API
-(`POST /api/admin/reports/schedules` with `"kind": "patrol_digest"`): the
-existing scheduler, cadence, recipients, and Pro reporting entitlement are
-reused, and the run renders the same digest as plain-language HTML and text
-(`internal/api/patrol_digest_email.go`) through the tenant email config via
-`SendEmailWithRetry`. Digest schedules are weekly and email-only; nothing is
-written to disk. A run without an email destination fails with a message that
-names the missing setting instead of silently doing nothing. Webhook and
-Apprise channels are out of scope.
-
-In Settings > Reporting, "Create schedule" offers a "Report type" choice
-between the performance report and the Patrol weekly summary. The choice only
-appears while the session reports that Patrol can run (AI on with a provider
-configured, the same capability that shows the Patrol navigation), and an
-existing summary schedule keeps showing its type so it can be edited after AI
-is switched off. Choosing the summary reduces the form to a name, weekday,
-time, timezone, and optional recipients: cadence is pinned to weekly, delivery
-to email, and the format, resources, tag filter, retention, attachment, and
-save-copy controls are hidden because the server fixes those values. The
-schedules table describes a summary's scope as "Patrol activity, last 7 days".
-
-## API
-
-`GET /api/ai/patrol/digest` — requires `ai:execute` scope like the other
-Patrol read endpoints. Query `days` (1 to 30, default 7). Returns the window,
-current mode, and the run, finding, investigation, action, alert, and spend
-rollups described above. Empty history returns zero counts with
-`window.history_complete = true`; a missing AI or Patrol service returns the
-same zero shape so the card can render its "Patrol has not run" state.
+For setup, interpretation, email checks and retained-data limits, see
+[Weekly summary and email](AI.md#weekly-summary-and-email) in the Pulse
+Intelligence guide. The summary uses existing Patrol activity; it is not a
+complete audit or proof that every resource is healthy.
diff --git a/frontend-modern/public/docs/AI.md b/frontend-modern/public/docs/AI.md
index 5ee09e222f..bce42d5bd5 100644
--- a/frontend-modern/public/docs/AI.md
+++ b/frontend-modern/public/docs/AI.md
@@ -494,6 +494,122 @@ When control level is **Controlled**, write actions pause for approval:
---
+## Weekly summary and email
+
+Pulse's **This week** summary adds up retained Patrol activity over the last
+seven days. Open **Patrol → Activity** to read it; use **Settings → Reporting**
+to schedule the same summary by email. The card, weekly email and **Report type**
+selector are included in **v6.5.0**. An older build can lack these controls.
+
+This is a summary of recorded activity, not a complete audit or proof that every
+resource is healthy. Use individual findings, run history and **Actions** for
+the result of a particular investigation or fix.
+
+### Read the summary
+
+The card normally says **Last 7 days**: a rolling window, not a calendar week.
+If it says **Since … (older runs are no longer kept)**, its run history covers
+only part of that window. Refreshing the card reads existing records; it does
+not run Patrol or recover missing history.
+
+| Line | What it means and what to check |
+| --- | --- |
+| **Patrol runs** | Retained runs, checks and failures in the window. The resource count is the largest number checked in one run, not a count of distinct resources across the week. Read run history for failed or missing checks. |
+| **New issues** | New-finding totals recorded by those runs. The open severity counts concern findings detected within the window; older open findings are not included. Check the full findings list before concluding that nothing needs attention. |
+| **Issues resolved** | Recorded automatic resolutions plus manual resolutions still retained. Dismissed and muted findings are described separately. A cleared finding does not establish that an action ran or was independently verified. |
+| **Investigated** | Recorded investigation outcomes on retained findings. “Fix run, result unknown” is not a verified fix; open the finding for its result and remaining work. |
+| **Fixes run** | Actions recorded as successfully executed. “Verified afterwards” is a separate count; review failed actions and waiting approvals in **Actions**. Do not run an action again merely to increase the count. |
+| **Estimated spend** | Recorded Patrol model calls and their estimated USD cost, not an invoice or all Assistant usage. If some calls have no known price, the displayed amount is incomplete, not a zero-cost result. Check [Cost Tracking](AI.md#cost-tracking). |
+
+The displayed mode is Patrol's **current** mode, not a history of mode changes.
+Watch-only mode can explain why investigations or fixes were not performed,
+but changing mode is not a repair for an empty summary.
+
+### Schedule the weekly email
+
+Scheduling requires the **advanced reporting** entitlement and an account
+allowed to manage report schedules. Reading the Patrol card does not itself
+grant scheduling access. See [Plans and entitlements](PULSE_PRO.md).
+
+1. Check that email notifications are **enabled and configured** under
+ **Alerts → Notifications**. Listing recipients on a schedule does not
+ configure a mail provider or enable email.
+2. Open **Settings → Reporting**, choose **Create schedule**, then set
+ **Report type** to **Patrol weekly summary**.
+3. Enter a schedule name, weekday, time and an IANA **Timezone**, such as
+ `Europe/London`. The summary is weekly, covers the last seven days, and is
+ sent as HTML and plain text email. It is not a PDF/CSV attachment, a
+ resource/tag-filtered report, or a saved disk copy.
+4. Set **Email recipients** deliberately. Blank uses the existing email
+ notification recipients; if none are configured, delivery can fall back
+ to the sender address. Use explicit authorised recipients to avoid that
+ ambiguity. The email can contain resource names, findings and costs.
+5. Save with **Enabled** selected for automatic runs. **Run now** sends a real
+ email; it is not a preview or a test of future scheduled delivery. Use it
+ only when those recipients should receive the summary.
+
+The Report type choice appears when the session reports AI as enabled. An
+existing summary keeps its type when edited after AI is switched off. If the
+choice is missing, check the installed version, current workspace, AI settings
+and reporting access; do not change Patrol's autonomy mode, re-enrol agents or
+create an API token just to reveal it.
+
+### Check a missing email or empty summary
+
+- **No recorded runs:** check Patrol's enabled state, configured provider,
+ ordinary run history and errors. “Patrol has not run” can also reflect an
+ unavailable Patrol service, not proof that the environment had no problems.
+ Do not force a run, restart Pulse or change retention just to fill the card.
+- **Failed schedule:** read **Last run** and the saved error in Reporting. On
+ a narrow screen, use the same signed-in page at a wider viewport to see
+ that column. A missing Patrol service and missing email configuration are
+ different failures. Preserve the original error; do not repeatedly use
+ Run now or send synthetic alerts to diagnose them.
+- **Run marked OK but no email received:** that status means Pulse's sender
+ returned success, not that the message reached the recipient's inbox.
+ Check the selected recipients and the mail provider's delivery or rejection
+ record, then the mailbox's filtering. A successful alert Test does not prove
+ that this schedule ran or used the intended recipients.
+- **Email links missing:** links depend on the configured Pulse public URL.
+ Open Patrol in your existing authenticated browser; do not make Pulse
+ publicly accessible or disable TLS to make an email link work.
+
+Keep full mail errors, recipient addresses, credentials and exports private.
+Share only a locally reviewed, redacted error and the relevant schedule time
+and outcome; see [Getting Help](TROUBLESHOOTING.md#-getting-help).
+
+### Honest limits of the data
+
+- Run history retains at most **100 runs**. Event-triggered activity can
+ exhaust it before seven days. The partial-window notice concerns run
+ retention only; its absence does not certify all other lines as complete.
+- Resolved findings can be purged after **24 hours**, and dismissed findings
+ after **30 days**. Manual resolutions, dismissals and investigation outcomes
+ on purged findings can disappear from these totals. Automatic resolution
+ totals come from retained runs, so the lines need not reconcile exactly.
+- Actions have their own retained audit records, but the digest reads a bounded
+ set and can show an empty action line when that source is unavailable.
+ Check **Actions** itself; zero in this summary is not proof that no action
+ took place. Pending approvals can include actions created before the window.
+- Model spend depends on recorded usage and known prices. This summary cannot
+ establish that caching, a budget limit or a provider's bill is correct.
+
+### API
+
+For an existing authenticated integration, `GET /api/ai/patrol/digest` requires
+`ai:execute` scope. Optional `days` is an integer from **1 to 30**, default
+**7**; invalid values return 400. Scheduled email always uses seven days.
+The response groups `window`, `mode`, `runs`, `findings`, `investigations`,
+`actions`, `alerts` and `spend`. `window.history_complete` and
+`window.history_since` describe the run-retention limitation above, not a
+whole-report completeness guarantee.
+
+A zero response can be returned when the Patrol service is unavailable.
+Authentication, scope or transport errors are not a zero result. Prefer the
+signed-in card for a one-off check; never extract a session cookie or put an
+API token in a command, URL or public report. See [API authentication](API.md#-authentication)
+for the private header-file approach when an integration needs it.
+
## Configuration
Configure providers in the UI: **Settings → Pulse Intelligence → Provider & Models**
diff --git a/frontend-modern/src/utils/__tests__/patrolDigestHelp.test.ts b/frontend-modern/src/utils/__tests__/patrolDigestHelp.test.ts
new file mode 100644
index 0000000000..3f8901a2b8
--- /dev/null
+++ b/frontend-modern/src/utils/__tests__/patrolDigestHelp.test.ts
@@ -0,0 +1,90 @@
+import { readFileSync } from 'node:fs';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { describe, expect, it } from 'vitest';
+
+import { renderDocMarkdown } from '@/features/docs/docMarkdown';
+
+const root = path.resolve(path.dirname(fileURLToPath(import.meta.url)), '../../../..');
+const read = (relative: string) => readFileSync(path.join(root, relative), 'utf8');
+const markdown = read('docs/AI.md');
+const article = document.createElement('article');
+const rendered = document.createElement('article');
+rendered.innerHTML = renderDocMarkdown(markdown, 'AI');
+for (
+ let element = rendered.querySelector('#weekly-summary-and-email');
+ element;
+ element = element.nextElementSibling
+) {
+ if (element.tagName === 'H2' && element.id !== 'weekly-summary-and-email') break;
+ article.append(element.cloneNode(true));
+}
+const text = article.textContent!.replace(/\s+/g, ' ');
+
+// Exercise the actual shipped Markdown, not a second copy of the advice.
+describe('Patrol weekly summary help', () => {
+ it('ships operator help instead of a private build or sales plan', () => {
+ expect(read('frontend-modern/public/docs/AI.md')).toBe(markdown);
+ expect(text).toContain('included in v6.5.0');
+ expect(text).not.toMatch(/Status: building|past due|stops paying|FEATURE_REQUESTS/);
+ expect(read('docs/PATROL_WEEKLY_DIGEST.md')).toContain('(AI.md#weekly-summary-and-email)');
+ expect(article.querySelector('#read-the-summary')).not.toBeNull();
+ expect(article.querySelector('#check-a-missing-email-or-empty-summary')).not.toBeNull();
+ });
+
+ it('makes enabled mail and reporting access prerequisites of the setup flow', () => {
+ const steps = Array.from(article.querySelectorAll('ol li'), (item) =>
+ item.textContent!.replace(/\s+/g, ' '),
+ );
+ expect(steps).toHaveLength(5);
+ expect(text).toContain('advanced reporting entitlement');
+ expect(steps[0]).toContain(
+ 'Listing recipients on a schedule does not configure a mail provider or enable email',
+ );
+ expect(steps[1]).toContain('Report type to Patrol weekly summary');
+ expect(steps[2]).toContain('Timezone');
+ expect(steps[2]).toContain('not a PDF/CSV attachment');
+ });
+
+ it('warns about recipient fallback and real Run now delivery', () => {
+ expect(text).toContain('delivery can fall back to the sender address');
+ expect(text).toContain('Run now sends a real email; it is not a preview');
+ expect(text).toContain("Pulse's sender returned success");
+ });
+
+ it('does not turn run or action counters into whole-estate acceptance', () => {
+ expect(text).toContain('largest number checked in one run, not a count of distinct resources');
+ expect(text).toContain('older open findings are not included');
+ expect(text).toContain('current mode, not a history of mode changes');
+ expect(text).toContain('zero in this summary is not proof that no action took place');
+ expect(text).toContain('Pending approvals can include actions created before the window');
+ });
+
+ it('separates missing inputs, partial history, unknown spend and inbox receipt', () => {
+ expect(text).toContain('partial-window notice concerns run retention only');
+ expect(text).toContain('displayed amount is incomplete, not a zero-cost result');
+ expect(text).toContain('not that the message reached the recipient');
+ expect(text).toContain('unavailable Patrol service, not proof');
+ expect(text).toContain(
+ 'Manual resolutions, dismissals and investigation outcomes on purged findings',
+ );
+ });
+
+ it('preserves API errors and routes private evidence to real help anchors', () => {
+ expect(text).toContain('ai:execute scope');
+ expect(text).toContain('invalid values return 400');
+ expect(text).toContain('Authentication, scope or transport errors are not a zero result');
+ expect(text).toContain('never extract a session cookie');
+ for (const [slug, fragment] of [
+ ['AI', 'cost-tracking'],
+ ['API', '-authentication'],
+ ['TROUBLESHOOTING', '-getting-help'],
+ ]) {
+ const link = article.querySelector(`a[href="/docs/${slug}#${fragment}"]`);
+ expect(link?.hasAttribute('data-doc-link')).toBe(true);
+ const target = document.createElement('article');
+ target.innerHTML = renderDocMarkdown(read(`docs/${slug}.md`), slug);
+ expect(target.querySelector(`[id="${fragment}"]`)).not.toBeNull();
+ }
+ });
+});