菜单
Codex reviews a plan. Returns dense verdict + feedback (not prose). Caller specifies reasoning effort (medium/high/xhigh) based on review depth needed.
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
Review Qodo findings on the open GitHub pull request for the current branch using the persistent review as the canonical finding list and GitHub inline threads as discussion anchors, then fix, verify, refresh, or escalate with `/ask` using evidence and Qodo acknowledgment state.
Delegate planning to Codex/GPT. Returns dense, AI-optimized plan (not prose). Caller specifies reasoning effort (high/xhigh) based on complexity.
Semantic version control workflow: sem (structural truth) -> inspect (review prioritization) -> weave (coordination + merge semantics).
Use the Google Workspace CLI to work with Gmail, Drive, Calendar, Docs, and Sheets from the local host shell.
Guide for creating or updating agent skills. Use when users want to create a new skill, update an existing skill, or need guidance on SKILL.md structure and best practices across Claude Code, Codex, and other compatible agents.
| name | codex-review |
| description | Codex reviews a plan. Returns dense verdict + feedback (not prose). Caller specifies reasoning effort (medium/high/xhigh) based on review depth needed. |
| context | fork |
| model | haiku |
| allowed-tools | ["Bash","Read","Glob","Grep"] |
Delegate plan review to Codex/GPT-5 for critical second opinion. Return response verbatim.
The caller (main agent) should specify:
medium, high, or xhigh based on review depth
medium: Standard plan validationhigh: Complex plans, architectural decisionsxhigh: Security-critical, novel systems, high-stakes reviewsmedium)codex exec -m gpt-5.2-codex -c model_reasoning_effort="[medium|high|xhigh]" --sandbox read-only -C "[PROJECT_DIR]" "[PROMPT]"
1. TASK: Review [plan] for completeness and correctness.
2. EXPECTED OUTCOME: Verdict + specific feedback in dense format.
3. CONTEXT:
Plan:
[PLAN CONTENT]
Goals: [objectives]
Constraints: [limits if known]
4. CONSTRAINTS:
- Evaluate against actual codebase
- Be practical, not theoretical
- Consider existing patterns
5. MUST DO:
- Evaluate: clarity, completeness, architecture, verifiability
- Simulate doing the work to find gaps
- Check for over-engineering
- Identify missing edge cases
- Provide specific improvements if rejecting
6. MUST NOT DO:
- Write or edit files
- Rubber-stamp approval
- Give vague feedback
7. OUTPUT FORMAT:
Dense, AI-agent-optimized format. No prose.
VERDICT: [APPROVE | REJECT | APPROVE_WITH_CONCERNS]
SCORES:
- Clarity: [1-5]
- Completeness: [1-5]
- Architecture: [1-5]
- Verifiability: [1-5]
CONCERNS:
- [issue]: [why it matters]
REQUIRED_CHANGES: [if REJECT]
- [change]
SUGGESTIONS: [optional improvements]
- [suggestion]
Return Codex response verbatim. Format is intentionally terse for token efficiency.