菜单
Validate implementation plans against requirements and design. Use after creating a plan to ensure full coverage and no ambiguities.
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
Structure Claude prompts for clarity and better results using roles, explicit instructions, context, positive framing, and strategic organization. Use when crafting prompts for complex tasks, long documents, tool workflows, or code generation.
Pause agent execution to have the human validate decisions, choose between options, answer freetext, or comment on a document — via the `hl` CLI. Every interaction is a kickoff that returns a job handle immediately; collect the human's answer later with `hl job result`. Use for material design decisions, approval gates, picks between meaningful alternatives, and markdown doc review. Not for trivial yes/no confirmations the agent should decide itself.
Validate technical designs against requirements for coverage, feasibility, and architectural soundness. Use after creating a design.
Validate requirements documents for EARS compliance, completeness, and consistency. Use after creating requirements.
Design CLIs for humans and LLM agents — subcommand shape, output streams, exit codes, JSON modes, TTY-aware color, structured errors, mutation safety. Use when building or refactoring a CLI, adding machine-readable output, writing --help, deciding stdout vs stderr, or making a tool agent-friendly.
Systematic debugging methodologies for hard bugs. Covers rubber ducking, code tracing, hypothesis testing, logging strategies. Use when stuck on bugs that resist quick fixes.
| name | review-plan |
| description | Validate implementation plans against requirements and design. Use after creating a plan to ensure full coverage and no ambiguities. |
| user-invocable | false |
| allowed-tools | Task |
Input: $ARGUMENTS
Parse input to extract requirements path, design path, and plan path. Expected format: {requirements-path} {design-path} {plan-path} or topic name (resolve from .claude/specs/{topic}/ and .claude/plans/).
Spawn a Task agent to review:
Task tool parameters:
subagent_type: general-purpose
model: sonnet
description: Review plan against requirements and design
prompt: |
Review plan against requirements and design.
Paths from input: $ARGUMENTS
- Requirements: `.claude/specs/{topic}/requirements.md`
- Design: `.claude/specs/{topic}/design.md`
- Plan: `.claude/plans/*.plan.md`
## Process
1. Read all three documents (requirements first, then design, then plan)
2. Extract every acceptance criterion from requirements
3. Extract every architectural decision from design
4. Map each acceptance criterion to plan coverage: covered, partial, or missing
5. Verify plan implements the design's architecture (not an alternative approach)
6. Check plan quality
**Threshold: Only flag issues that would block implementation or cause genuine confusion.**
### Quality Checks
- Ambiguous language → only if implementation would stall or go wrong
- Deferred decisions → only if missing info needed to start work
- Conditional branches → only if unresolved decisions (valid branching is fine)
- Unexplored complexity → only if it hides surprising work
- Design divergence → plan contradicts design's architecture without justification
## Output
If all covered and no issues: `PASS`
If issues exist:
1. Missing: [acceptance criterion from requirements not in plan]
2. Design mismatch: [plan diverges from design without justification]
3. Ambiguous: [unclear section that needs resolution]
Report the agent's findings to the user.