agentleFS
Sign inStart free
RoyoTechRoyoTech/royo-learn/skills/publish-learning/SKILL.md

publish-learning

Skill7 starsChanged 4 months agoMITView on GitHub

Publish only a curated and approved learning. Generate and show a preview first, obtain human approval for AGENTS.md, shared libraries or existing Skill changes, apply through royo-learn, verify, audit and rollback on failure.

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 8 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈8

    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. Very specific:

    When the preview reports requires_approval: true (AGENTS.md, shared scope or existing Skill updates), obtain explicit human approval, then …

    line 27
  2. Very specific:

    Call learning_publish with the exact preview hash, and pass the approval_id returned by learning_approve whenever approval was required.

    line 31
  3. Distinctive:

    Verify result and registry.

    line 33
  4. Distinctive:

    If verification fails, ensure rollback occurred.

    line 34
  5. Distinctive:

    never treat a preview as an authorization: a preview describes, it does not permit.

    line 49
  6. Very specific:

    never reuse an approval for a different preview — learning_publish will reject it.

    line 51
  7. Very specific:

    never edit targets directly when royo-learn can apply the approved operation.

    line 56
  8. Distinctive:

    never reuse approval for another preview.

    line 57
  9. Distinctive:

    never place a long procedure in AGENTS.md.

    line 58
  10. Distinctive:

    never guess a global path.

    line 59
  11. Distinctive:

    never bypass a dirty-target block silently.

    line 60
  12. Distinctive:

    never publish secrets or project-private assumptions into shared scope.

    line 61

≈ “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

Publish Learning

Preconditions

  • status is approved;
  • curation specifies target and action;
  • acceptance checks exist;
  • no unresolved conflict;
  • preview is current;
  • required approval exists.

Workflow

  1. Call learning_publication_preview.
  2. Inspect targets, diff, risk and verification.
  3. Show high-impact previews to the user.
  4. When the preview reports requires_approval: true (AGENTS.md, shared scope or existing Skill updates), obtain explicit human approval, then call learning_approve with the exact preview hash to record it. See "Approval step" below.
  5. Call learning_publish with the exact preview hash, and pass the approval_id returned by learning_approve whenever approval was required.
  6. Verify result and registry.
  7. If verification fails, ensure rollback occurred.
  8. Report canonical target and publication ID.

Approval step

royo-learn now enforces approval. When a preview reports requires_approval: true, publishing is refused until a valid approval exists, so the burden is shared with the tool rather than resting entirely on you.

  • obtain explicit human approval in conversation first;
  • record it with learning_approve, passing the exact preview_hash, approved_by, reason and approval_evidence (a link, message id or ticket);
  • the approval is bound to that preview hash. It is rejected if the preview changes, a destination changes, the prior file content of a destination changes, the relevant policy changes, it expires, or it is revoked;
  • never treat a preview as an authorization: a preview describes, it does not permit;
  • never reuse an approval for a different preview — learning_publish will reject it.

Rules

  • never edit targets directly when royo-learn can apply the approved operation;
  • never reuse approval for another preview;
  • never place a long procedure in AGENTS.md;
  • never guess a global path;
  • never bypass a dirty-target block silently;
  • never publish secrets or project-private assumptions into shared scope.

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