[assafbar2]

What kind of change?

• New agent listing (content/agents/<slug>.json added)
• Correction to an existing entry (URL fix, renamed product, new access method…)
• Removal (product discontinued)
• Code / site change

Summary

Adds nine tasks to content/tasks.json, 47 in all. They are defined like the existing ones (label, blurb, categories, access, keywords) and placed next to related tasks, after the first eight that the homepage chips show.

Task	Fits
analyze-product-usage	9
query-bi-dashboards	13
manage-data-pipelines	13
run-ci-cd-pipelines	10
test-web-and-mobile-apps	7
secure-identity-and-access	7
create-and-sign-documents	3
draw-diagrams-and-3d	8
build-websites-and-internal-apps	11

The tasks fit 76 listings. 27 of them are from the Nov 6 drop and fit no task until now; 49 were already in the directory (PostHog, Tableau, CircleCI, BrowserStack, DocuSign, Miro, Webflow and others). Six one-off listings fit none of these and stay without a task: Granola, Instacart, OpenJarvis, RTK, Tigris and Zillow.

Other changes:

• Aperture by Tailscale: its description now says it authenticates users by their tailnet identity, as Tailscale's docs say ("uses Tailscale's identity layer to automatically authenticate users"). This is what lets it fit the identity task. The edit was made through scripts/cms.ts, so it is logged in content/changelog.json.
• Keyword choices: keywords were picked to avoid noise. For example, "lineage" is out because the matcher reads typed "linear" as the same word, which would stop "linear" from routing to manage tickets; "data lineage" is used instead.
• Routing test: lib/task-match.test.ts now expects "build an agent" to tie three tasks, since the new "build websites and internal apps" label shares the word "build". It still picks no task.
• Docs: the /api/tasks example in docs/API.md now shows a total of 47.
• Release: 2.9.0, with a CHANGELOG entry whose Highlights link to the new task pages.

Test timeouts

The catalog-wide unit tests walk every listing (996) and every task. On shared CI runners several of them already took 3–4 s per case against vitest's 5 s default (for example, the sitemap listing-dates test ran 4.0–4.2 s on recent PR branches). Nine more tasks pushed two of them over: the sponsored ranking test at 5.8 s and the sitemap test at 5.0 s. None of the assertions change.

• vitest.config.ts sets testTimeout: 15_000.
• tests/sitemap.test.ts builds the sitemap once in beforeAll instead of once per case.
• lib/sponsored.test.ts checks one task per case, with the alternatives and Operator checks as their own cases.

Checklist

• npx tsx scripts/validate-content.ts passes locally (47 tasks, 0 violations)
• Added a line to content/changelog.json
• npm test (760 passed), typecheck and lint (no new warnings) pass locally
• Listing and catalog changes are contributed under CC BY-NC 4.0 (see LICENSE-DATA); code under MIT