[17.4.0] - 2026-09-16 - "YouTube Transcript Skill, SEO-AEO Audit-First Overhaul, and Catalog Persistence"
Adds one reviewed community skill, rewrites the SEO-AEO family around an audit-first evidence workflow, and hardens catalog persistence.
A maintenance release for Claude Code, Cursor, Codex CLI and Gemini CLI.
It adds one community skill and one new SEO-AEO skill reviewed through the
protected maintainer workflow, rewrites the SEO-AEO skill family around an
audit-first evidence workflow, and keeps the hosted catalog aligned with actual
browser storage; the catalog contains 2,124 skills and existing installation
interfaces remain supported.
Start here
- Install:
npx agentic-awesome-skills@17.4.0 - Choose your tool
- Best skills by tool
- Bundles
- Workflows
Added
youtube-transcript-skills: fetch YouTube transcripts, search videos and
channels, browse channels, and extract playlists via the
getyoutubetranscript.com API.seo-aeo-orchestrator: end-to-end SEO/AEO growth orchestrator that sequences
audit, keyword, content, internal-linking, schema, and meta outputs.
Changed
- Rewrote ten
seo-aeo-*skills around an audit-first, evidence-first growth
workflow with explicit anti-fabrication and "Not assessed" rules.
Fixed
- Mark catalog skills as saved only after local persistence succeeds.
- Keep the catalog shortlist aligned with the last confirmed browser write.
- Ignore stale catalog refresh results in the web app.
Improved
- Refreshed canonical indexes, catalogs, plugin mirrors, editorial bundles, and
web assets for the 2,124-skill release. - Allow catalog web-app browser source through the fork-safety intake.
- Allow the pinned AAS agent-first preview workflow in fork-run approvals.
- Official Atlas Cloud badge and sponsorship attribution in README/docs.
Validation scope
The source PRs passed maintainer repair, exact-head review, repository
validation, reference and documentation-security checks, protected CI, CodeQL,
dependency review, package publication and release-only Pages verification.
Thanks to @pushkarsingh32 (#1434), @jhuang-tt (#1446), @jhuang-tt (#1459),
@ShaunLinTW (#1448), and @prewsh (#1460) for their contributions.