Skip to main content
ClaudeWave
Skill1.1k repo starsupdated 3d ago

skill-lifecycle-promotion

Use when adding or auditing Agentlas skill lifecycle metadata, skill-registry.json, trial evidence, Curator promotion decisions, or first-class skill promotion gates.

Install in Claude Code
Copy
git clone --depth 1 https://github.com/agentlas-ai/Agentlas-OS /tmp/skill-lifecycle-promotion && cp -r /tmp/skill-lifecycle-promotion/.agents/skills/skill-lifecycle-promotion ~/.claude/skills/skill-lifecycle-promotion
Then start a new Claude Code session; the skill loads automatically.

SKILL.md

# Skill Lifecycle Promotion

Use this skill when a generated or packaged Agentlas repo needs governed skill
promotion metadata.

## Procedure

1. Add `.agentlas/skill-registry.json` as an export-only candidate registry.
2. Add empty `.agentlas/skill-trials.jsonl` and
   `.agentlas/curator-decisions.jsonl`.
3. Keep every skill at `tier: candidate` on export.
4. Keep `runtimeFirstClassRecallEnabled: false` unless a local Curator later
   approves promotion.
5. Add success predicates and situation tags for every skill.
6. Separate `## Memory Events` from `## Skill Trial Events`.
7. Treat LLM rubric review as weak evidence only.
8. Block promotion when authority separation, sealed holdouts, replayability, or
   rollback evidence is missing.
9. Include false-accept, blind-spot, and drift terms in any durable-error budget.

## Output

Return:

- registry files added or checked;
- promotion tier status;
- evidence gaps;
- rollback/quarantine status;
- residual risks.