跳到正文
cccskills
免费在 GitHub 公开

plan

Plan a non-trivial task before implementing. Use when the task requires 3+ steps, multiple files, or an architectural decision. Writes a checkable plan to claude/tasks/todo.md.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(1)

  • SKILL.md2.0 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Plan Skill

Create a structured, checkable implementation plan before writing any code.

When to use

  • Task requires 3 or more steps
  • Task touches multiple files or layers
  • Task involves an architectural decision
  • Any time the path forward is not immediately obvious

Protocol

1. Read mandatory context first

  • claude/tasks/todo.md — current task state (are we resuming?)
  • claude/tasks/lessons.md — relevant past lessons
  • claude/architecture.md — project structure
  • claude/rules.md — golden rules

2. Understand the full scope

  • What is the exact goal? What does done look like?
  • What layers does this touch (schema, service, API, tests, docs)?
  • What already exists that is relevant?
  • What are the failure modes?

3. Write the plan to claude/tasks/todo.md

Format:

# Task: <Clear, concise title>

**Goal:** <One sentence. What is done when this is complete?>
**Branch:** <git branch name, if applicable>
**Loaded docs:** <which claude/*.md files you read>

## Steps
- [ ] 1. <First action — specific file or command>
- [ ] 2. <Second action>
- [ ] 3. <Third action>
...

## Review Checklist
- [ ] Tests pass
- [ ] Linter passes
- [ ] No new warnings
- [ ] PR description ready (if applicable)

## Notes
<Any constraints, risks, or decisions to track>

4. Validate the plan (optional — for complex tasks)

Spawn a plan-challenger subagent:

"Spawn plan-challenger to review this plan. Here is the plan: [paste plan]"

5. Get approval for non-trivial plans

Present the plan to the user. Wait for "proceed" or revisions before implementing.

Rules

  • NEVER start implementing before the plan is written
  • NEVER mark a step complete without proof (test result, output, etc.)
  • If something goes wrong mid-execution: STOP, re-read the plan, update it, then continue

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

ask

免费

Route a codebase question to the right tool — structural graph, semantic search, or risk analysis. Use when you need to find or understand code without knowing exact file names.

中文简介正在准备中。当前显示原文说明。

brain-bootstrap/claude-code-brain-bootstrap112026年4月25日 更新

bootstrap

免费

Auto-configure CLAUDE.md and claude/ knowledge docs for a new project. Use when setting up Brain Bootstrap in a new repository. Discovers the codebase, fills in templates, and writes project-specific configuration.

中文简介正在准备中。当前显示原文说明。

brain-bootstrap/claude-code-brain-bootstrap112026年4月25日 更新

Use before any new feature, component, or behavioral change — when you need to explore requirements and design before writing code

中文简介正在准备中。当前显示原文说明。

brain-bootstrap/claude-code-brain-bootstrap112026年4月25日 更新

careful

免费

Activate safety guards for the session. Use /careful before critical operations — blocks rm -rf, DROP TABLE, git push --force, kubectl delete namespace, docker system prune via PreToolUse hook.

中文简介正在准备中。当前显示原文说明。

brain-bootstrap/claude-code-brain-bootstrap112026年4月25日 更新

changelog

免费

Generate a user-facing changelog from git commits — categorize, filter noise, translate to user language. Runs in isolated fork context.

中文简介正在准备中。当前显示原文说明。

brain-bootstrap/claude-code-brain-bootstrap112026年4月25日 更新

Save session state before context gets full or before ending. Writes current task state, branch, and loaded docs to claude/tasks/todo.md so the next session can resume cleanly.

中文简介正在准备中。当前显示原文说明。

brain-bootstrap/claude-code-brain-bootstrap112026年4月25日 更新

查看 brain-bootstrap 的全部技能

报告此技能的问题