agentleFS
Sign inStart free
Zengxiansheng007Zengxiansheng007/skills-old/skills/rag-system/SKILL.md

rag-system

Skill1 starsChanged 3 months agoMITView on GitHub

Use when the user asks to use the local D:/RAG knowledge base, query RC/evidence, ingest user-provided materials, prepare downstream packets for auto-testcase-generator/ui-test/Solution D, validate RAG governance, or route to rag-* child skills.

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 2 are distinctive, and 11 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 11

    Very specific

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

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

    Treat D:/RAG as a controlled local evidence store, not a free disk search surface.

    line 10
  2. Very specific:

    Read only user-listed files or files already registered in D:/RAG/00_inbox/candidate-materials.json.

    line 11
  3. Distinctive:

    Do not ingest secrets, credentials, tokens, cookies, or unredacted production/customer data.

    line 12
  4. Very specific:

    Never promote active-candidate, active, uncertain, or conflict material to formal RC.

    line 13
  5. Very specific:

    Do not execute ui-test, create formal test cases, or write downstream systems from this skill.

    line 15
  6. Distinctive:

    Read references/routing-policy.md for the current request type.

    line 19
  7. Very specific:

    If the request is ingestion or file admission, hand off to rag-intake.

    line 21
  8. Very specific:

    If the request is schema, packet, or fixture validation, hand off to rag-schema.

    line 22
  9. Very specific:

    If the request is RC, conflict, or confirmation governance, hand off to rag-governance.

    line 23
  10. Very specific:

    If the request is evidence lookup or citation preparation, hand off to rag-query.

    line 24
  11. Very specific:

    If the request is source-version or impact analysis, hand off to rag-change-impact.

    line 25
  12. Very specific:

    If the request is for auto-testcase-generator, ui-test, or Solution D, hand off to rag-downstream-handoff.

    line 26
  13. Very specific:

    Stop and explain the blocked reason if a request would expose sensitive data, read unlisted files, use active as RC, cross a knowledge …

    line 50

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Rag System

Operating Rules

  • Treat D:/RAG as a controlled local evidence store, not a free disk search surface.
  • Read only user-listed files or files already registered in D:/RAG/00_inbox/candidate-materials.json.
  • Do not ingest secrets, credentials, tokens, cookies, or unredacted production/customer data.
  • Never promote active-candidate, active, uncertain, or conflict material to formal RC.
  • Formal downstream answers require confirmed_active RC or an explicit user decision recorded in governance.
  • Do not execute ui-test, create formal test cases, or write downstream systems from this skill. Produce read-only packets only.

Workflow

  1. Read references/routing-policy.md for the current request type.
  2. Load D:/RAG/manifest.json and verify the request has a known knowledge_space_id.
  3. If the request is ingestion or file admission, hand off to rag-intake.
  4. If the request is schema, packet, or fixture validation, hand off to rag-schema.
  5. If the request is RC, conflict, or confirmation governance, hand off to rag-governance.
  6. If the request is evidence lookup or citation preparation, hand off to rag-query.
  7. If the request is source-version or impact analysis, hand off to rag-change-impact.
  8. If the request is for auto-testcase-generator, ui-test, or Solution D, hand off to rag-downstream-handoff.
  9. Return a concise result with status, evidence refs, blocked reasons, and next required user decisions.

Decision Rules

  • Missing or mismatched knowledge_space_id: stop before reading material content.
  • Registered blocked-pending-sensitive-review: refuse ingestion, summary, RC, and downstream use.
  • Registered active-candidate: allow candidate packets and review notes only.
  • Conflicts: return need-review and link the conflict record.
  • Requests to scan a directory: require a user-provided file list first.

Validation

Run:

python D:\RAG\scripts\validate_rag_mvp.py
python C:\Users\lenovo\.codex\skills\rag-system\scripts\forward_test_rag_system.py

Also validate this skill folder with the local skill validator when available.

Safety

Stop and explain the blocked reason if a request would expose sensitive data, read unlisted files, use active as RC, cross a knowledge space boundary, or perform downstream execution.

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.

Check your setup, free