agentleFS
curv-institutecurv-institute/atlas/.claude/skills/sync-repo-knowledge/SKILL.md

sync-repo-knowledge

Skill0 starsChanged 4 months agoMITView on GitHub

Read a repo and populate this hub's wiki with a knowledge page about it (purpose, structure, key docs, current state, recent activity, open items). Use when the user says "/sync-repo-knowledge <path>", "sync knowledge from <repo>", or "refresh what Atlas knows about <repo>". Idempotent: re-running refreshes the generated section and leaves hand-edited notes intact. /add-repo calls this after registering a repo.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read a git checkout and write or refresh a topic page at wiki/topics/<repo-name>.md summarizing what this hub should know about it.

    line 8
  2. Very specific:

    Link it: add the page under Topics in Home.md, and set the repo's REPOS.md row Role/notes to point at wiki/topics/<name>.md, if not already …

    line 31
  3. Very specific:

    Keep it factual and concise.

    line 33

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

/sync-repo-knowledge <path> - populate wiki knowledge from a repo

Read a git checkout and write or refresh a topic page at wiki/topics/<repo-name>.md summarizing what this hub should know about it. Idempotent and non-destructive: regenerate only the marked generated zone; never overwrite hand-edited notes. This summarizes a repo, it does not vendor it in.

Steps

  1. Resolve <path> (default: current directory). Confirm it is a git checkout (git -C <path> rev-parse --is-inside-work-tree); if not, report and stop.
  2. name = basename of git -C <path> rev-parse --show-toplevel (the same canonical basename rule as /add-repo, so the page filename matches the REPOS row exactly). Target page: wiki/topics/<name>.md.
  3. Gather knowledge (read-only; do not modify the source repo):
    • Purpose: the first paragraph of the repo README (README.md / .rst / docs index).
    • Structure: top-level dirs and their role (one line each); key entry points.
    • Canonical state: any CURRENT_STATE / STATUS / CHANGELOG / ROADMAP / docs index in the repo.
    • Notable docs: AGENTS.md / CONTRIBUTING / design notes (one line each).
    • Recent activity: git -C <path> log -10 --format='%h %ad %s' --date=short.
    • Open items: counts of TODO / FIXME / XXX; any open-questions or issues docs.
    • Provenance: remote (git remote get-url origin), default branch, current HEAD commit.
  4. Write the page in two zones:
    • A generated zone between <!-- atlas:generated:start --> and <!-- atlas:generated:end --> holding the gathered knowledge plus a provenance line (repo path, remote, synced-at HEAD). On re-run, replace ONLY this zone, leaving everything else untouched.
    • A ## Notes (hand-edited) heading below the markers, created once and never overwritten.
  5. Link it: add the page under Topics in Home.md, and set the repo's REPOS.md row Role/notes to point at wiki/topics/<name>.md, if not already linked.
  6. Keep it factual and concise; obey STYLE.md (no AI tells). Report what was written. Commit only if asked.

Notes

  • Provenance: the generated zone records the source HEAD so staleness is visible. Re-run to refresh.
  • Privacy: for a public Atlas instance, do not paste sensitive internal content into the wiki. Summarize at a safe level, or keep the page in a gitignored location.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.