agentleFS
Sign inStart free
Gaurav890Gaurav890/everything-agentic-engineering/.claude/skills/harness-evolution/SKILL.md

harness-evolution

Skill1 starsChanged 2 months agoMITView on GitHub

Evaluate bounded harness candidates against the protected incumbent using sanitized outcome signals and deterministic quality, regression, safety, cost, and latency gates. Use when project evidence suggests instructions, examples, memory curation, or routing should improve; never use it to self-promote, change protected evals, expand authority, or train model weights.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    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. Distinctive:

    Use the repository command surface:

    line 8
  2. Very specific:

    Start with low-risk few-shot examples.

    line 23

≈ “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

Harness evolution

Use the repository command surface:

./agentic evolve status
./agentic evolve signal validate <sanitized-signal.json>
./agentic evolve compare --candidate <candidate-results.json>

Follow:

ACT → SANITIZE SIGNAL → EVALUATE → PROPOSE → COMPARE → HUMAN REVIEW → RECORD

Read docs/30-engineering/HARNESS_EVOLUTION.md and docs/60-tooling/EVOLUTION_POLICY.md before preparing a candidate.

Start with low-risk few-shot examples. Instructions, memory curation, and routing require the reviews declared in .agentic/evolution/policy.json. Protected policy, evaluation, security, workflow, dependency, credential, permission, tool, and production surfaces are not candidate-owned.

The committed engine is offline and proposal-only. It may validate sanitized signals and compare previously produced evaluation results. It must not ingest raw production traces, invoke a remote model, implement its own proposal, change its own exam, install or enable capabilities, deploy, canary, promote, approve, or merge.

PASS means only that a candidate is eligible for independent human review. It is never promotion authority.

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