agentleFS
Sign inStart free
loganware05loganware05/captains-compass-cursor/.cursor/skills/experience-routing/SKILL.md

experience-routing

Skill2 starsChanged 32 days agoMITView on GitHub

Proposes Skill confidence and matcher weight adjustments from Experiences (proposal-only until Captain-flagged apply)

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 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.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Do not auto-apply matcher weights from proposal generation

    line 64
  2. Very specific:

    Do not mark proficiency captain_approved: true without Captain confirmation

    line 65
  3. Very specific:

    Do not advance candidates past SANDBOX_TESTED

    line 66
  4. Distinctive:

    Do not land persistent roles without a Captain-reviewed PR

    line 67

≈ “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

Experience Routing

Use this Skill when

Closing a workstream with Experiences available, or when the Captain asks for routing improvements based on historical outcomes.

Inputs

  • One or more Experience JSON files (fixtures or .agent/experience/)
  • Optional notes

Procedure

  1. Collect Experience paths (prefer committed fixtures for reproducible proposals).

  2. Generate a proposal-only routing artifact:

    ./scripts/propose-experience-routing.sh \
  --experiences tests/fixtures/experience/contact-counter.json
  3. Confirm auto_apply is false and captain_approved is false by default.
  4. Present skill confidence deltas and weight suggestions to the Captain.
  5. To apply weights (Milestone 4 Level 3), Captain sets captain_approved: true on the proposal JSON, then use Skill bounded-autonomy / ./scripts/apply-routing-proposal.sh under an autonomy budget. Routing proposals include decomposition hints (bounded matcher-factor deltas; M17) merged on apply.

  6. For context selection tuning (M17 — which knowledge/TI slices appear in plans):

    ./scripts/propose-context-selection.sh --experiences tests/fixtures/experience/contact-counter.json

    Captain sets captain_approved: true, then ./scripts/apply-context-selection-proposal.sh.

  7. Optionally record subagent proficiency metadata after Skill training:

    ./scripts/record-agent-proficiency.sh \
  --agent-id compass-evaluator \
  --classifications evaluation,arbitration \
  --skills compass-evaluator,execution-telemetry \
  --level developing \
  --captain-approved false

    Set --captain-approved true only after explicit Captain approval.

  8. For persistent specialist roles, use Skill persistent-role-promotion (staging + PR only).

Output

  • Routing proposal under .agent/routing/proposals/
  • Optional proficiency draft under .agent/agents/proficiency/

Prohibited actions

  • Do not auto-apply matcher weights from proposal generation
  • Do not mark proficiency captain_approved: true without Captain confirmation
  • Do not advance candidates past SANDBOX_TESTED
  • Do not land persistent roles without a Captain-reviewed PR

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