agentleFS
jsvillalbatjsvillalbat/sevito-skills/skills/compound-knowledge/SKILL.md

compound-knowledge

Skill0 starsChanged 2 months agoMITView on GitHub

Captures a solved non-trivial problem as a searchable markdown doc in docs/solutions/ (root cause, fix, prevention, keywords), and searches that folder before starting new non-trivial debugging so past solutions get reused instead of re-derived from scratch. Use right after fixing a tricky bug, working around a gotcha, or solving something that took real investigation — and before diagnosing a new problem that might already have a documented answer.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. JairoTorregrosa/jaiskills★ 11same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository8 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

7 rules. About 0 are common, 3 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    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:

    Every hard-won fix that only lives in one session's context gets re-derived from scratch the next time it comes up — in a different …

    line 11
  2. Very specific:

    Search: before starting non-trivial debugging or investigation — check docs/solutions/ first in case this was already solved.

    line 26
  3. Very specific:

    Confirm it's actually solved: Don't write up a fix you're only hoping works — verify it per verify-before-done's evidence ladder first.

    line 33
  4. Very specific:

    Check for a near-duplicate first: Search docs/solutions/ by keyword/category — if this updates or extends an existing doc rather than being …

    line 35
  5. Very specific:

    Before diving into a new non-trivial problem, grep/glob docs/solutions/ for matching keywords, error text, or category.

    line 60
  6. Distinctive:

    If a match exists, read it fully before starting fresh investigation — apply it directly if the situation matches, or note explicitly why …

    line 62
  7. Distinctive:

    If no match exists, proceed normally — and capture the result per the steps above once solved.

    line 65

≈ “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 Knowledge

Every hard-won fix that only lives in one session's context gets re-derived from scratch the next time it comes up — in a different session, by a different agent, or by the same agent six months later with no memory of last time. This skill turns a solved problem into a searchable artifact so the investigation compounds instead of repeating.

When to use this

  • Capture right after closing out a problem that took genuine investigation: a non-obvious root cause, a workaround for a library/platform quirk, a gotcha that will plausibly bite again. Not "fixed a typo" — that doesn't need a doc.
  • Triggered by the user saying "that worked," "remember this," or on your own initiative right after verifying a hard-won fix (see verify-before-done — don't document a fix you haven't actually confirmed works).
  • Search before starting non-trivial debugging or investigation — check docs/solutions/ first in case this was already solved.

Instructions

Capturing a solution

  1. Confirm it's actually solved. Don't write up a fix you're only hoping works — verify it per verify-before-done's evidence ladder first.
  2. Check for a near-duplicate first. Search docs/solutions/ by keyword/category — if this updates or extends an existing doc rather than being genuinely new, update that file instead of creating a near-duplicate.
  3. Pick a category from a small fixed set so the folder stays scannable: build-errors, runtime-bugs, performance, security, integration-config, tooling. If nothing fits, it's fine to add a new category — just keep the set small.
  4. Write one file at docs/solutions/<category>/YYYY-MM-DD-<kebab-description>.md using the template in references/template.md. Cover, in your own single pass (no need for separate sub-analyses — one thorough pass covering all three angles is enough):
    • Problem — the symptom, how it showed up, what it looked like from the outside.
    • Root cause — what was actually wrong, not just what fixed it.
    • Solution — what changed, with file:line references where useful.
    • Prevention — how to catch this earlier next time: a test to add, a check, a convention.
    • Keywords — terms someone would actually search for when hitting this again, including error message fragments.
  5. Consider a targeted follow-up check proportional to the category — a security fix might warrant checking for the same pattern elsewhere in the codebase; a performance fix might warrant a benchmark; a build fix might warrant confirming the full test suite still runs clean.

Searching before investigating

  1. Before diving into a new non-trivial problem, grep/glob docs/solutions/ for matching keywords, error text, or category.
  2. If a match exists, read it fully before starting fresh investigation — apply it directly if the situation matches, or note explicitly why this case differs if it doesn't quite fit.
  3. If no match exists, proceed normally — and capture the result per the steps above once solved.

Reference files

  • references/template.md — the blank solution-doc template.
  • references/example.md — one filled-in worked example.

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.