AGENT SKILL

skill-publish

Publish a skill after validation and approvals; updates the catalog and logs the publish event.

License 许可
AGPL-3.0
最近更新
2026-08-15
Stars
★ 30
Forks
7

01 前言 Front Matter

name
skill-publish
description
Publish a skill after validation and approvals; updates the catalog and logs the publish event.
allowed-tools
Read, Grep, Glob, ApplyPatch, Write
argument-hint
[skill-name]
disable-model-invocation
true

02 资源文件 1

03 SKILL.md 正文

Skill Publish — Controlled Release

Purpose

Publish a validated skill with proper approvals, then update catalog metadata and log the publish event.

Inputs

  • Skill name

  • Evidence of test gates

  • Explicit approver (required for high blast radius)

  • Template paths are resolved from ${AGENTS_SKILLS_ROOT}/skill-publish/.

Steps

  1. Run /skill-validate <skill-name> and confirm PASS.
  2. Enforce approval gate
    • If blast_radius: high, require explicit approval from an authorized maintainer. If missing, STOP.
  3. Update catalog entry
    • Increment version if needed
    • Update updated date
  4. Write publish log to .jarvis/context/skills/publish-log.md with:
    • Skill name, version, owner, blast radius
    • Approver (if required)
    • Date and summary
  5. Return confirmation

Decision Trace Protocol

This skill participates in the skill evolution system by capturing decision traces at gate resolutions and consulting accumulated feedback.

Trace Consultation (short loop)

Before executing any step with a quality or human gate, query accumulated decision traces:

python3 "${AGENTS_SKILLS_ROOT}/_shared/trace_query.py" recent \
    --skill "skill-publish" --gate "<gate_name>" --limit 5 --min-loops 1

If patterns or recent feedback exist, incorporate them as additional constraints. Do not cite traces to the user unless asked.

Trace Capture (after gate resolution)

After every gate resolves, capture a decision trace:

python3 "${AGENTS_SKILLS_ROOT}/_shared/trace_capture.py" capture \
    --skill "skill-publish" \
    --gate "<gate_name>" --gate-type "<human|quality>" \
    --outcome "<approved|rejected|passed|failed>" \
    --refinement-loops <N> \
    [--feedback "<user's feedback text>"] \
    [--category <CATEGORY>]

Post-Run Retrospective

After completion, ask: "Any feedback on this skill-publish run? (skip to finish)" If provided, capture via trace_capture.py with gate "run_retrospective" and gate-type "retrospective".

Output

  • Confirmation message
  • Link to catalog entry + publish log

04 同仓库更多 Skills