Skill Proposal: [Short Title]

Source

• URL:
• Title:
• Author or organization:
• Source type:
• Retrieval status:
• Evaluation file:
• Decision: APPROVE / HUMAN_REVIEW / REJECT

Mechanism

Describe the implementable mechanism in 3-6 sentences. Include what changes in an agent's behavior, what context or harness surface is affected, and what evidence supports it.

Skill Target

• Target type: new skill / existing skill / reference only
• Target path:
• Activation scenario:
• Related skills:
• Proposed location: SKILL.md / references / scripts / researcher-only

Novelty Check

• Command:
• Verdict: pass / humanreview / likelyduplicate
• Max mechanism overlap:
• Top mechanism overlaps:
• Human-review rationale:

Evidence

Only include evidence from retrieved sources. If a source was discovered but not fetched, record it under gaps instead of citing it.

Proposed Delta

changes:
  - path: ""
    section: ""
    change_type: "add | update | remove"
    summary: ""

Quality Checks

• [ ] Fits an existing activation scenario or justifies a new one.
• [ ] Adds an operating rule, workflow, artifact, gotcha, or reference.
• [ ] Records novelty-check verdict and top mechanism overlaps.
• [ ] Avoids duplicating existing skill guidance or accepted mechanisms.
• [ ] Keeps SKILL.md under 500 lines.
• [ ] Uses progressive disclosure for detailed or volatile evidence.
• [ ] Uses platform-agnostic wording.
• [ ] Updates README, root SKILL.md, and manifests if publishing a new skill.

Risks And Gaps

• Evidence limitations:
• Possible duplication:
• Volatile claims:
• Required human review:

Recommendation

State one of:

• publish-new-skill
• update-existing-skill
• add-reference-only
• human-review
• reject

Then provide a short rationale focused on why this change should or should not enter the published corpus.