From request to release
Four stages, and today two of them are done by a person reading your issue — not a pipeline. Here is exactly what happens when you propose a skill for this catalog.
-
1.0Open a request
Open a Skill Request issue on GitHub. It asks for the skill name, a one-paragraph purpose, why it's empowering, a suggested category and tier, and any similar skills already in the catalog.
-
2.0Review
The repo owner reads the issue and checks it against the quality bar above. This is a person reading and deciding, today — not an automated triage step.
-
3.0Build
If it's approved, the owner or a contributor writes the actual skill files —
SKILL.md,README.md, and theskills-index.jsonentry — and opens a pull request. Someone builds it; nothing generates it on its own yet. -
4.0Merge & live
CI validates the frontmatter, checks that
skills-index.jsonis in sync, and confirms the per-skill page exists. Once the owner merges, GitHub Pages rebuilds from the merged branch on its own — that part really is automatic, and the skill is live in the catalog.
What's automated today, and what isn't
Stage 4's rebuild is automatic — GitHub Pages redeploys from main without anyone touching a button. Stages 2 and 3 are manual: a person reviews the issue, and a person (or a contributor) writes the files and opens the pull request. A faster, more automated review-to-PR pipeline is on the roadmap, but it is not built yet — nothing here files a PR for you.
Pending external sources
Separately from the request pipeline above, Owen sometimes finds skill-like content in other public repos and proposes adapting it into this catalog. Each one below is awaiting his individual review and approval — none of it is part of the catalog yet, and none of it gets the "adapted from / credited to" treatment an approved skill gets until that review happens.
- Loading pending sources…
Ready to propose one?
Start at stage one — the issue template asks for everything the review needs.
Start at stage 1 →