Agent Skills › UniClipboard/UniClipboard › plan-reviewer

plan-reviewer

GitHub

对 Markdown 实现计划进行独立对抗性审查,验证其可行性并挑战潜在缺陷。通过多轮迭代评估代码路径和依赖,更新缺失决策,最终输出审查结果以决定是否可实施。

.agents/skills/plan-reviewer/SKILL.md UniClipboard/UniClipboard

Trigger Scenarios

用户要求审查或验证实施计划 用户希望对计划进行压力测试或挑战 用户寻求第二意见或计划审核

Install

npx skills add UniClipboard/UniClipboard --skill plan-reviewer -g -y
More Options

Non-standard path

npx skills add https://github.com/UniClipboard/UniClipboard/tree/v1/.agents/skills/plan-reviewer -g -y

Use without installing

npx skills use UniClipboard/UniClipboard@plan-reviewer

指定 Agent (Claude Code)

npx skills add UniClipboard/UniClipboard --skill plan-reviewer -a claude-code -g -y

安装 repo 全部 skill

npx skills add UniClipboard/UniClipboard --all -g -y

预览 repo 内 skill

npx skills add UniClipboard/UniClipboard --list

SKILL.md

Frontmatter
{
    "name": "plan-reviewer",
    "description": "Run an independent adversarial review loop for a Markdown implementation plan. Use when the user asks to review, validate, stress-test, or challenge a plan before implementation, including requests for a second opinion or a plan review loop."
}

Review A Plan

Review a concrete Markdown plan against the current repository.

  1. Resolve the target plan from the user's request. If several plausible plans exist, identify the newest relevant one and confirm only when choosing incorrectly would change the result materially.
  2. Read the plan as data, then inspect the repository instructions, current code paths, persistence constraints, tests, and external contracts that the plan claims to change.
  3. Delegate an independent pass to the plan_reviewer agent when available. Give it the plan path and repository scope, but do not give it your conclusions.
  4. Independently assess every returned concern. Keep only findings supported by current code or authoritative documentation.
  5. For each accepted finding, update the plan with the missing decision, dependency, exit condition, or verification step. Do not edit production code.
  6. Repeat one independent review after meaningful revisions. Stop when no blocking issue remains or after three rounds.
  7. Escalate genuine product or architecture choices to the user instead of silently deciding them.

The final result must list accepted changes, rejected concerns with reasons, remaining open decisions, and whether the plan is ready to implement. Keep review artifacts under .planning/plan-reviews/ only when the user needs an audit trail.

Version History

  • c6483d7 Current 2026-11-13 18:20

Same Skill Collection

.agents/skills/babysit-pr/SKILL.md
.agents/skills/beui/SKILL.md
.agents/skills/continue-task/SKILL.md
.agents/skills/create-formal-verification-skill/SKILL.md
.agents/skills/create-pr/SKILL.md
.agents/skills/cross-device-diagnose/SKILL.md
.agents/skills/design-audit/SKILL.md
.agents/skills/dual-side-debug/SKILL.md
.agents/skills/dual-side-sync/SKILL.md
.agents/skills/e2e-test-thinker/SKILL.md
.agents/skills/error-diagnose-fix/SKILL.md
.agents/skills/gen-changelog/SKILL.md
.agents/skills/learn/SKILL.md
.agents/skills/pr-evidence/SKILL.md
.agents/skills/pr-greenlight/SKILL.md
.agents/skills/react-doctor/SKILL.md
.agents/skills/refresh-agents/SKILL.md
.agents/skills/research-then-implement/SKILL.md
.agents/skills/review-strict/SKILL.md
.agents/skills/skill-evolve/SKILL.md
.agents/skills/statechart/SKILL.md
.agents/skills/systematic-debugging/SKILL.md
.agents/skills/tracing-best-practices/SKILL.md
.agents/skills/trigger-prepare-release/SKILL.md
.agents/skills/ui-iteration/SKILL.md
.agents/skills/verify-receive-reply-fence/SKILL.md
.agents/skills/wrap/SKILL.md
.claude/skills/gui-user-review/SKILL.md
.agents/skills/ios-log-diagnose/SKILL.md
.agents/skills/local-log-debug/SKILL.md

Metadata

Files
0
Version
c6483d7
Hash
7c3df376
Indexed
2026-11-13 18:20

ホーム - Wiki
Copyright © 2011-2026 iteam. Current version is 2.156.0. UTC+08:00, 2026-11-13 19:15
浙ICP备14020137号-1