agentleFS
aegnticaegntic/compound-engineering/skills/compound-docs/SKILL.md

compound-docs

Skill1 starsChanged 5 months agoMITView on GitHub

Capture solved problems as categorized documentation with YAML frontmatter for fast lookup

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

13 rules. About 0 are common, 9 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If critical context is missing, stop and ask for it before writing anything.

    line 32
  2. Very specific:

    Search docs/solutions/ for similar documents.

    line 33
  3. Very specific:

    Create the solution doc with assets/resolution-template.md, then offer the post-capture menu.

    line 36
  4. Very specific:

    Use category mapping from references/yaml-schema.md to place the file under docs/solutions/<category>/.

    line 48
  5. Distinctive:

    After writing the document, present a terse next-step menu:

    line 53
  6. Very specific:

    Add to Required Reading using assets/critical-pattern-template.md.

    line 55
  7. Distinctive:

    Add the learning to an existing skill.

    line 57
  8. Distinctive:

    Create a new skill seeded by this solution.

    line 58
  9. Distinctive:

    Do not proceed when module, symptom, stage, or resolution steps are missing.

    line 68
  10. Distinctive:

    Do not skip schema validation.

    line 69
  11. Distinctive:

    Do not write vague titles, vague symptoms, or root causes like "fixed code".

    line 70
  12. Distinctive:

    Do not auto-promote to critical patterns.

    line 71
  13. Distinctive:

    Prefer one high-signal document over scattered notes across multiple files.

    line 72

≈ “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

compound-docs Skill

Capture non-trivial solved problems as searchable documentation in docs/solutions/. Keep the file short, structured, and schema-valid so future sessions can actually find and trust it.

When to use

  • A bug, migration issue, incident, or tricky setup problem is confirmed fixed.
  • The solution required real investigation, multiple attempts, or a non-obvious insight.
  • The team would benefit from preserving the failure mode, root cause, and prevention guidance.
  • Skip trivial typos, obvious syntax mistakes, and fixes that needed no meaningful investigation.

Workflow

  1. Detect a confirmed fix. Auto-invoke on clear success phrases such as "that worked", "it's fixed", or "problem solved", or run manually.
  2. Gather the required context: module, exact symptom/error, failed attempts, root cause, fix, prevention guidance, file references, and relevant environment details.
  3. If critical context is missing, stop and ask for it before writing anything.
  4. Search docs/solutions/ for similar documents. Reuse or cross-link only when the root cause genuinely matches.
  5. Build a filename as [sanitized-symptom]-[module]-[YYYYMMDD].md.
  6. Validate YAML frontmatter against schema.yaml using the enum guidance in references/yaml-schema.md.
  7. Create the solution doc with assets/resolution-template.md, then offer the post-capture menu.

Required context

  • Module or subsystem name.
  • Exact symptom or error message.
  • Stage or execution context.
  • What failed during investigation.
  • Root cause and what changed to fix it.
  • Prevention guidance and any follow-up actions.

Validation rules

  • problem_type, component, root_cause, and severity must match the schema exactly.
  • Use category mapping from references/yaml-schema.md to place the file under docs/solutions/<category>/.
  • Block on invalid YAML, missing required fields, or vague symptoms.
  • Include concrete examples or commands whenever the fix would be hard to repeat without them.

Post-capture menu

After writing the document, present a terse next-step menu:

  1. Continue workflow.
  2. Add to Required Reading using assets/critical-pattern-template.md.
  3. Link related issues.
  4. Add the learning to an existing skill.
  5. Create a new skill seeded by this solution.
  6. View the captured document.
  7. Other.

Output

  • A validated file in docs/solutions/<category>/[filename].md.
  • Cross-references when similar issues exist.
  • A short summary of what was captured and the next-step menu.

Guardrails

  • Do not proceed when module, symptom, stage, or resolution steps are missing.
  • Do not skip schema validation.
  • Do not write vague titles, vague symptoms, or root causes like "fixed code".
  • Do not auto-promote to critical patterns; offer the option and let the user choose.
  • Prefer one high-signal document over scattered notes across multiple files.

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.