跳到正文
cccskills
免费在 GitHub 公开

pb-release-aos

Release a new AOS version to npm through the release-please PR: check the commit subjects, review and gate the release PR, merge it only after GO, watch the release workflow and check version drift. Use for "release AOS", "cut an AOS release", "merge the release PR", "publish a new AOS version".

中文简介正在准备中。当前显示原文说明。

集成工具GitHub
查看安装方法

包含的文件(1)

  • SKILL.md5.2 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Release AOS to npm

What you get: a new AOS version on npm through the release-please PR, never a hand bump, with the PR reviewed and gated before your GO.

Inputs

  • repo (optional) — the AOS checkout, default the current directory; <owner/repo> comes from git -C <repo> remote get-url origin

Steps

  1. Preflight — gh auth status and bdb-ecosystem-health listed in the harness skill list or its SKILL.md under the harness skills dir (~/.claude/skills, ~/.agents/skills, ~/.codex/skills, ~/.config/opencode/skills) → run log header — gh missing or unauthenticated → log it, give the gh auth login hint, stop; not installed → log "health skill not installed" and skip the drift report in step 7; the npm check still runs
  2. github — commits since the last tag (git -C <repo> fetch --tags, then git -C <repo> describe --tags --abbrev=0, then git -C <repo> log <tag>..HEAD --format=%s) → subject list in the run log, every non-Conventional subject flagged (release-please cannot see it, see AGENTS.md) — stops for approval
  3. github — gh pr list -R <owner/repo> --state open --head release-please--branches--main --json number,title,headRefName → the release PR number in the run log — none → stop with "no release PR"
  4. Review and gate on that PR, using these skills:
    • git-pr-review — gh pr view <n> -R <owner/repo> --json commits → description draft in the run log — draft only, nothing posted
    • pr-recap — recap file written → link it; recap not buildable → log "pr-recap skipped: <reason>" and paste gh pr diff <n> -R <owner/repo> --name-only instead; no npx without approval
    • reviewer (agent; no subagent support → the main session reads agents/reviewer.md and runs the review inline, using only the diff and the contract, never the session's reasoning) — gh pr diff <n> -R <owner/repo>; the contract is the commit list from step 2, never the PR body → findings table — any open blocking finding stops the run
    • bdb-shipping-skill — door class of the release PR (two-way or one-way; unclear counts as one-way) → class in the run log; one-way → ADR-lite production_artifacts/decisions/<date>-<slug>.md — class recorded
    • godmode-shipping — gh pr checks <n> -R <owner/repo> all pass (pending, gh pr checks exit 8, counts as not passed), plus the local gate exactly as pb-ship step 8 gives it (scratch worktree from pull/<n>/head, lint, typecheck and test scripts that exist, a fork PR skips the local gate and is held, worktree removed without --force) → exit codes in the run log
  5. [GO] github — gh pr merge <n> -R <owner/repo> --squash — the WAITING FOR GO line names PR number, title, base branch, method and the version in .release-please-manifest.json. The run stops here until the human types GO. (Where the go-gate hook runs — Claude Code, OpenCode, agy — it covers gh pr merge; otherwise this GO is the only guard.) Never npm version or npm publish by hand.
  6. github — <sha> = the merge commit (gh pr view <n> -R <owner/repo> --json mergeCommit -q .mergeCommit.oid); gh run list -R <owner/repo> --workflow release-please.yml --commit <sha> --limit 1 --json databaseId (wait until a run exists), then gh run watch <id> -R <owner/repo> --exit-status → conclusion in the run log — exit 0
  7. npm check and drift — npm view @hybridlabor-api/aos version → run log — always run; equals the version in .release-please-manifest.json on main. bdb-ecosystem-health drift report only if step 1 found the skill
  8. quick-recap — run log → final line 🟢|🟡|🔴 with the released version — line written

Run log: production_artifacts/pb-release-aos-<date>.md in the start directory, never committed

Rules

  • Anything other than the literal GO (case-insensitive) is not a GO; a GO covers only that one step, one time.
  • A failed check stops the run: write the failure into the run log and report. No silent retries.
  • Write one run-log line per step as it completes (N. done|skipped|failed — artifact — check result) and WAITING FOR GO: <step> at each gate.
  • GO is by contract in every harness. The go-gate hook is only a backstop on Claude Code, OpenCode and agy (Codex: unverified; Cursor, Kimi: none). A missing hook is never permission to proceed.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Use when manage multiple local CLI agents via tmux sessions (start/stop/monitor/assign) with cron-friendly scheduling.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Use when a hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Reference for how BDB structures autonomous software engineering work — the seven-node dispatcher graph (Architect, TechLead, UI/UX, Engineering, Media/EventTech, Reviewer, Shipping) that /startcycle-graph actually runs. Use when you need the high-level lifecycle framing without inventing your own process.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Tools are how AI agents interact with the world. A well-designed tool is the difference between an agent that works and one that hallucinates, fails silently, or costs 10x more tokens than necessary. This skill covers tool design from schema to error handling.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

Harness patterns for coding agents — memory, permissions, context engineering, delegation, skills, hooks, bootstrap.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

agenttrail: live map of a multi-agent build in the browser: which plan component is being worked on, by which agent or harness, what is done and what is stuck. Use when a multi-agent pipeline starts (/startcycle, /startcycle-graph, /teamwork-preview) or after a plan-canvas approve, when the user asks to see what the agents are doing ("live map", "build board", "who is running now"), or by yourself whenever a multi-agent run is under way — no user prompt needed.

中文简介正在准备中。当前显示原文说明。

hybridlabor-api/aos62026年10月16日 更新

查看 hybridlabor-api 的全部技能

报告此技能的问题