Skip to content

Reposition as the open, rated index + ship browsable Top-100 into the repo - #3

Open
MARUCIE wants to merge 2 commits into
mainfrom
data-native-index
Open

Reposition as the open, rated index + ship browsable Top-100 into the repo#3
MARUCIE wants to merge 2 commits into
mainfrom
data-native-index

Conversation

@MARUCIE

@MARUCIE MARUCIE commented May 30, 2026

Copy link
Copy Markdown
Owner

Why

Stars go to repos that are themselves the destination (awesome-list pattern), not to a website's backend. The 40MB unified-index.json already lives in the repo but nobody bookmarks a JSON blob — the value was present but not presented.

What

  • data/TOP-SKILLS.md — browsable, rated Top-100 table (the bookmark hook)
  • scripts/generate-top-skills.py — regenerates it; every count is reproducible from the index meta block, so badges can't silently drift
  • README first-screen reposition — leads with "the open, rated index of AI agent skills", surfaces the data above the fold, demotes the live site + self-host to secondary, differentiates vs awesome-mcp-servers (unrated)
  • Fixes: count drift 37,296 → 37,310 (meta truth); removes an internal "HITL-tracked separately" note that leaked into the public README

Note

Merging triggers the CF production deploy (deploy.yml on push to main) — held as a PR so you control deploy timing. Docs/data only; no app code touched.

Maurice Wen added 2 commits May 30, 2026 22:57
Turn the 40MB unified-index.json into something a developer stars: a readable,
rated table of the top S-tier skills. Generated by scripts/generate-top-skills.py
so every count (37,310 total / 2,049 S-tier / source split) is reproducible from
the index meta block, not hand-typed — drift-proof against the README badges.
…screen

- Lead with 'the open, rated index of AI agent skills', browse data right here
- Surface data/TOP-SKILLS.md (Top 100) + unified-index.json above the fold
- Differentiate vs awesome-mcp-servers (unrated) and the official MCP registry
- Demote the live site + Cloudflare self-host to secondary (data is the product)
- Fix count drift 37,296 -> 37,310 (meta source of truth); add reproducibility note
- Remove internal 'HITL-tracked separately' note leaking into the public README
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant