agentleFS
paslavskyipaslavskyi/llm-wiki/.claude/skills/kb-recall/SKILL.md

kb-recall

Skill1 starsChanged 33 days agoMITView on GitHub

Use to find existing product knowledge in this repo before answering, deciding, or capturing — navigates index-first (MAP → domain index → specific notes) instead of reading everything.

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 index/MAP.md to pick the relevant domain(s) by note counts and names.

    line 11
  2. Very specific:

    Follow links (and index/backlinks.json for reverse links) to pull in directly related notes — one hop unless the question needs more.

    line 15
  3. Very specific:

    Never read the whole knowledge/ tree.

    line 32

≈ “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

kb-recall — index-first retrieval

Goal: find the relevant notes for a question while reading as little as possible.

Steps

  1. Read index/MAP.md to pick the relevant domain(s) by note counts and names.
  2. Open the relevant index/<domain>.index.md. Use the summary column to shortlist candidate ids — do NOT open notes yet.
  3. Open only the shortlisted note files.
  4. Follow links (and index/backlinks.json for reverse links) to pull in directly related notes — one hop unless the question needs more. Backlinks cover all reference fields (goal scope, plan lists, timeline sequence, sources, relations), so "which goal/plan mentions X" is answered by backlinks.json too. For MEANING ("what mitigates RISK-003?", "who does FEAT-002 serve?") read index/relations.json — typed edges with inverses.
  5. Structured questions over fields ("all must-requirements still draft", "what serves PER-002", "unvalidated assumptions") → run a saved query (npm run query -- --list, then npm run query -- <name> [--param X]) instead of opening notes one by one. If none fits and the question will recur, add a queries/<name>.yml (format in lib/query.mjs); for a one-off, write the YAML to a scratch file and run npm run query -- <path>.yml.
  6. Report findings:
    • debug: list the ids you read and their statuses.
    • autonomous: summarize what's already known in plain language, no ids.

Rules

  • Never read the whole knowledge/ tree. If the index can't answer it, widen the shortlist by one hop at a time.
  • Recall is read-only. To write, hand off to kb-capture.

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.