集成

Spec Plan Build Review

试用

Run a proportional delivery lifecycle for software or skill work: clarify scope, create a concise plan, implement, verify, review, and ship. Use when the use...

它能做什么

Use this skill to turn non-trivial work into a controlled delivery loop. Keep it proportional: tiny edits can skip straight to implementation and verification.

技能文档

Spec Plan Build Review

Use this skill to turn non-trivial work into a controlled delivery loop. Keep it proportional: tiny edits can skip straight to implementation and verification.

Lifecycle

  1. Spec

    • Identify the user-visible outcome, affected repo or package, and non-goals.
    • Read local context before deciding the implementation shape.
    • If requirements conflict, resolve from source evidence or ask one focused question.
  2. Plan

    • Make a short checklist only when it helps.
    • Keep one item in progress at a time.
    • Assign verification before shipping, not after.
  3. Build

    • Reuse existing patterns and helpers.
    • Keep changes tightly scoped to the stated outcome.
    • Avoid unrelated refactors, formatting churn, and broad dependency changes.
  4. Verify

    • Run the narrowest reliable local checks first.
    • Add or update tests when behavior changed.
    • Record any checks that could not run and why.
  5. Review

    • Inspect the diff as if reviewing a PR.
    • Lead with defects, missed tests, user-visible regressions, and release risks.
    • For risky work, use a fresh-context or high-risk review skill before shipping.
  6. Ship

    • Confirm a clean worktree except intended changes.
    • Commit with a specific message.
    • Push and verify remote CI.
    • Create release artifacts only after CI is green.

Proportionality Guide

  • tiny: one file or docs-only edit; implement and run a direct check.
  • normal: small feature or bug fix; plan, implement, test, review.
  • release: public package, ClawHub skill, security-sensitive change, or migration; run full lifecycle and remote CI.
  • high-risk: destructive operations, credentials, permissions, money, production, public policy; add fresh-context skeptical review.

Review Personas

Use these lenses during the review step:

  • code reviewer: correctness, maintainability, local conventions.
  • test engineer: missing coverage, flaky assumptions, reproducibility.
  • security reviewer: permissions, secrets, data exposure, supply chain, release surface.

If the runtime supports real subagents and the task is large enough, fan out independent review passes. Otherwise, apply the lenses sequentially in the main context and keep findings concrete.

Ship Checklist

  • Local tests or validators pass.
  • Public metadata is current.
  • No generated temp files are staged.
  • GitHub Actions are green for the pushed commit or tag.
  • ClawHub publish uses the same version as the repo release.
  • Final status names the shipped version, checks, and any residual risk.

相关技能

诊断生产力系统反复失效的根因,给出最小干预——容量测算、瓶颈定位、可靠的本地记录。

作者 Iván855 次安装69 星标

执行 Git 操作(提交、分支、合并、变基、冲突解决与恢复)时强制套用安全规则。

作者 Iván535 次安装31 星标

把自然语言描述转为结构化 JSON,并由 mcp-diagram-generator MCP 服务生成 Draw.io、Mermaid 或 Excalidraw 图表文件。

作者 nssa.io1.0k 次安装47 星标

在本地磁盘以分类纯 Markdown 文件保存需要长期留存的事实,与智能体内置记忆并存。

作者 Iván559 次安装18 星标

以 AI 机器人身份加入视频会议,提供语音、虚拟形象与屏幕共享四种模式。

作者 johnpatternai22 次安装8 星标

zack-dev-cm 的更多技能

浏览全部技能

OpenClaw CWS Publisher is a public ClawHub Chrome Web Store publisher skill. Use it when the user says "chrome web store publisher", "extension release publi...

作者 zack-dev-cm29 次安装3 星标

SOTA Agent is a public ClawHub SOTA-campaign skill for CV and DS work. Use it when the user says "sota agent", "state of the art benchmark scouting", or want...

作者 zack-dev-cm28 次安装2 星标

Use when studying or normalizing Chinese lesson transcript/subtitle inputs with bundled public lesson data, learner docs, local export bundles, and pilot-fir...

作者 zack-dev-cm22 次安装1 星标

OpenClaw YouTube Publisher is a public ClawHub YouTube publishing skill. Use it when the user says "youtube creator ops", "youtube publisher", "YouTube Studi...

作者 zack-dev-cm25 次安装2 星标

Build source-backed research claim ledgers from drafts, literature matrices, notes, citation lists, PDFs, source packets, or reviewer comments. Use when the...

作者 zack-dev-cm13 次安装

Install, wire, audit, and harden AgentMemory for AI coding agents and project workspaces. Use when Codex needs to adopt rohitg00/agentmemory, configure Codex...

作者 zack-dev-cm14 次安装