菜单
This skill should be used when the user asks to "write a plan", "create an implementation plan", "plan this feature", or has an approved design that needs to be broken into executable tasks for implementation.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | write-plan |
| description | This skill should be used when the user asks to "write a plan", "create an implementation plan", "plan this feature", or has an approved design that needs to be broken into executable tasks for implementation. |
| version | 0.1.0 |
| superpowers | writing-plans |
Generate a highly detailed, executable implementation plan tailored for a "junior engineer" with zero context. Every task must be self-contained and actionable.
Save the plan as: docs/plans/YYYY-MM-DD-<feature-name>.md
# Implementation Plan: <Feature Name>
**Date:** YYYY-MM-DD
**Design Doc:** <link or reference>
**Estimated Tasks:** N
## Overview
<1-2 sentence summary>
## Tasks
### Task 1: <Short description>
**File:** `path/to/file.ext`
**Time:** ~X minutes
**Steps:**
1. <Specific action>
2. <Specific action>
**Code:**
\`\`\`language
// Exact code to add/modify
\`\`\`
**Verify:**
\`\`\`bash
<command to verify task is complete>
\`\`\`
**Commit:** `<conventional commit message>`
---
Each task should be:
At the end of the plan, include:
## Progress Tracker
- [ ] Task 1: <description>
- [ ] Task 2: <description>
...
## Notes
<Any context for resuming work later>
docs/plans/Once the plan is complete, suggest:
"Plan ready. Invoke the execute-plan skill to start implementing tasks sequentially."
Audit and rewrite content to remove AI writing patterns ("AI-isms"). Use this skill when asked to "remove AI-isms," "clean up AI writing," "edit writing for AI patterns," "audit writing for AI tells," or "make this sound less like AI." Supports a detect-only mode, an edit-in-place mode for files, an optional voice profile (casual / professional / technical / warm / blunt), and an iterate-to-convergence pass.
Use this skill when reviewing production code for dead code, duplication, unnecessary complexity, and non-idiomatic patterns. Use after implementing a feature, after refactors, or whenever code quality feedback is requested.
Use this skill to review tests for false confidence, excessive mocking, meaningless assertions, and weak behavior coverage. Use after writing tests or when a suite passes but bugs still escape.
Multi-agent debate with visible transcripts where agents respond to each other. USE WHEN council, debate, perspectives, weigh options, deliberate, multiple viewpoints. Unlike RedTeam (adversarial), Council is collaborative-adversarial.
Summarize or extract text/transcripts from URLs, podcasts, and local files (great fallback for “transcribe this YouTube/video”).
Use this skill when the user asks to "commit", "make a commit", "commit my changes", "create a commit", or invokes /commit. Detects the active VCS (git, jj, or hg), inspects the working copy, drafts a commit message in the project's standard format (terse summary + detailed paragraph + test plan), and creates the commit. Honors staged-files-only when files are already staged in git. Handles jj-specific traps (interactive editors, conflict states, the operation log, the post-commit working copy gotcha) without dropping into TUIs that block non-interactive shells.