agentleFS
deanmikandeanmikan/ai-native-company-starter/.agents/skills/capture-learning/SKILL.md

capture-learning

Skill0 starsChanged 2 months agoMITView on GitHub

Record a non-obvious insight discovered during work as a dated learning in knowledge/learnings/ so it compounds instead of being lost. Use at the end of any task that surfaced a constraint, failed approach, vendor quirk, corrected assumption, or hard-to-find fact, or when asked to "note that", "remember that", "write down what we learned".

14 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 14 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

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

    Check knowledge/learnings/ and the relevant area page for the same insight.

    line 26
  2. Distinctive:

    Write the title as the statement of what is now known, for example "Learning: the payments provider settles weekly, not daily".

    line 30
  3. Distinctive:

    If the learning contradicts an approved page, do not edit that page.

    line 37

“Common” is measured: kept by at least 6 in 10 of the 14 edited versions of this file we could read. “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

Capture a learning

Insight that stays in a conversation is lost. This skill takes under five minutes and is part of finishing work.

Is it a learning?

Apply the counterfactual test: if this were never written down, would a future person repeat a mistake, rediscover a fact at real cost, or take on material risk? Yes means write it. It must also not already have a home. Examples: an API limit that shaped a design, a vendor process that differs from its documentation, a metric definition that turned out to be ambiguous, an approach that failed and why, a number that took an hour to find.

No, if it is task status (tracker), a record of what was done (Git history), a choice with alternatives (record-decision), or something an existing page already says (update that page).

Steps

  1. Check knowledge/learnings/ and the relevant area page for the same insight. If a page already covers it, update the page instead.
  2. Create knowledge/learnings/YYYY-MM-DD-short-name.md from templates/learning.md. Use today's date.
  3. Write the title as the statement of what is now known, for example "Learning: the payments provider settles weekly, not daily".
  4. Fill every section briefly. "Evidence" must contain at least one source, observation, or reproduction step. "Prevention" says what to do differently. "Promote to" names the page this should eventually live in.
  5. Set owner to the person who can verify it, status: draft, and review_by 60 days out.
  6. If the learning contradicts an approved page, do not edit that page. Mention the conflict in the learning and tell the owner.
  7. Run python3 scripts/lint.py.
  8. Include the learning in the same branch or pull request as the work that produced it. A learning filed later is usually never filed.

Report

One line in your final message: "Captured learning: knowledge/learnings/... (promote to areas/...)".

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.