메뉴
Review the current plan by auto-selecting appropriate reviewer agents based on project characteristics.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
| name | plan-review |
| description | Review the current plan by auto-selecting appropriate reviewer agents based on project characteristics. |
| allowed-tools | Bash(codex exec *) |
Review the latest plan file by analyzing project characteristics, selecting appropriate reviewer agents, and running them in parallel.
~/.claude/agents/| Argument | Required | Description |
|---|---|---|
| agent-name | No | Explicit agent name. Omit for auto-selection (recommended) |
ls -t ./plans/*.md 2>/dev/null | head -1
If no file is found, report an error and stop. Use the Read tool to load the detected plan file.
If an agent name is provided as an argument, use only that agent (manual mode).
If no argument is given (recommended), auto-select reviewers based on project signals.
Collect the following signals in parallel:
| Signal | Detection Method |
|---|---|
| Rust project | Cargo.toml exists, or *.rs files present |
| Go project | go.mod exists, or *.go files present |
| TypeScript project | tsconfig.json exists, or *.ts files present |
| codex CLI available | which codex succeeds |
| Refactoring-related | Plan body contains refactoring keywords (see below) |
| Test infrastructure | Test files, test config, or test directories exist |
Refactoring keywords (checked against plan body):
Test infrastructure detection — at least one primary signal:
*.test.ts, *.spec.ts, *.test.tsx, *.spec.tsx, *.test.js, *.spec.js, *_test.go, *_test.rsvitest.config.*, jest.config.*, playwright.config.*tests/, __tests__/, test/Based on collected signals, select reviewers:
| Condition | Reviewer to Launch |
|---|---|
| codex CLI is available | codex-reviewer |
Before launching, show the selection summary:
Project analysis:
- codex CLI: Y (available)
Launching reviewers: codex-reviewer
Launch all selected agents in parallel via the Agent tool.
Prompt passed to each agent:
Please review the following Plan file.
Based on your expertise, provide feedback on:
1. Technical accuracy
2. Potential issues and risks
3. Improvement suggestions
4. Overlooked considerations
---
Plan File: <path>
---
<content>
Important: Agent tool calls are independent — always invoke multiple agents in a single message for true parallelism.
Aggregate all reviewer results in this format:
=== Plan Review Results ===
--- codex-reviewer ---
[feedback from codex-reviewer]
=== Summary ===
Cross-cutting summary of all reviewer feedback, with high-severity issues listed first.
| Agent Name | Auto-Select Condition | Expertise |
|---|---|---|
| codex-reviewer | codex CLI is available | General architecture & design review |
| Situation | Response |
|---|---|
| No plan file found | Notify that the plans directory has no files |
| No matching reviewers in auto-select | List available agents and suggest manual selection |
| Specified agent not found | List available agents and report error |
| Some agents fail | Report successful results and note failures |
./plans (relative to project root)Publish agent research as a durable Markdown vault entry, optionally with a GitHub Gist share artifact.
Iteratively ask clarifying questions to enrich the current Plan before review.
Create, switch, list, and delete git worktrees with `git wt` (the git-wt CLI). Use when working on multiple branches in parallel, isolating an experiment in its own checkout, or whenever the user mentions worktrees or `git wt`.
Extract technical learnings from the session and save them to the learn repository.
Iteratively refine the current Plan by applying Codex review feedback until Codex has no more issues to raise.
Analyze session context and generate a targeted prompt for /compact to preserve important information.