AgentSkillsCN

make-plan

运行代码质量检查工具，包括 ruff、black 和 mypy。通过格式化代码、检查代码风格、进行类型检查，确保代码符合项目规范、修复格式问题或进行代码审查时使用。

SKILL.md
--- frontmatter
name: make-plan
description: Create an implementation plan with documentation discovery. Use for complex features or multi-phase tasks that need structured planning before execution.

Make Plan Skill

Create LLM-friendly implementation plans that can be executed in phases, using subagents for research and fact-gathering.

When to Use

  • •Complex features requiring multiple steps
  • •Tasks that need documentation research first
  • •Multi-file or multi-system changes
  • •Anything where "just winging it" would lead to invented APIs

How It Works

You are an ORCHESTRATOR. Create plans in phases that can be executed consecutively.

Delegation Model

  • •Use subagents for fact gathering: docs, examples, signatures, grep results
  • •Keep synthesis and plan authoring with the orchestrator
  • •If a subagent report is incomplete, re-check with targeted reads/greps

Subagent Reporting Contract (MANDATORY)

Each subagent response must include:

  1. •Sources consulted - files/URLs and what was read
  2. •Concrete findings - exact API names/signatures, exact file paths
  3. •Copy-ready snippet locations - example files/sections to copy
  4. •Confidence note + gaps - what might still be missing

Reject and redeploy the subagent if it reports conclusions without sources.

Plan Structure

Phase 0: Documentation Discovery (ALWAYS FIRST)

Before planning implementation, deploy "Documentation Discovery" subagents to:

  1. •Search for and read relevant documentation, examples, and existing patterns
  2. •Identify the actual APIs, methods, and signatures available (not assumed)
  3. •Create a brief "Allowed APIs" list citing specific documentation sources
  4. •Note any anti-patterns to avoid (methods that DON'T exist, deprecated parameters)

Then consolidate findings into a single Phase 0 output.

Each Implementation Phase Must Include

  1. •What to implement - Frame tasks to COPY from docs, not transform existing code
    • •✅ Good: "Copy the V2 session pattern from docs/examples.ts:45-60"
    • •❌ Bad: "Migrate the existing code to V2"
  2. •Documentation references - Cite specific files/lines for patterns to follow
  3. •Verification checklist - How to prove this phase worked (tests, grep checks)
  4. •Anti-pattern guards - What NOT to do (invented APIs, undocumented params)

Final Phase: Verification

  1. •Verify all implementations match documentation
  2. •Check for anti-patterns (grep for known bad patterns)
  3. •Run tests to confirm functionality

Key Principles

  • •Documentation Availability ≠ Usage: Explicitly require reading docs
  • •Task Framing Matters: Direct agents to docs, not just outcomes
  • •Verify > Assume: Require proof, not assumptions about APIs
  • •Session Boundaries: Each phase should be self-contained with its own doc references

Anti-Patterns to Prevent

  • •❌ Inventing API methods that "should" exist
  • •❌ Adding parameters not in documentation
  • •❌ Skipping verification steps
  • •❌ Assuming structure without checking examples

Example Usage

When asked to implement a complex feature:

code
User: Add WebSocket support to the server

You (orchestrator):
1. Deploy subagent: "Research WebSocket implementation patterns in this codebase and any docs"
2. Receive findings with specific file references and API signatures
3. Write Phase 0 output with allowed APIs
4. Write Phase 1-N implementation phases with verification checklists
5. Output the complete plan for execution with /do-plan

Output Format

Write the plan to a file (e.g., plans/feature-name.md) so it can be referenced by /do-plan.