法典

人工智能 社区

技能解读

设计思路

codex 是 gstack 给 Claude Code 接入 OpenAI Codex CLI 的桥梁——不是替代 Claude,而是让 Claude 在审视自己的工作时找一个独立的第二意见。skill 把命令封成 /codex review / /codex challenge / /codex consult 三个 verb,每次跑一个只读 sandbox 的 codex,让它单独审一遍代码或思路,然后把输出逐字搬回给用户看。

工作流

  • /codex review:让 codex 对当前 diff / 文件做独立 review,5 分钟外层超时,10 分钟内层 timeout wrapper。
  • /codex challenge:让 codex 唱反调,专门挑 Claude 可能漏掉的 edge case 和反例。
  • /codex consult:把它当 senior 资深顾问问问题。
  • 默认会 --web-search-cached 启用 OpenAI 的缓存检索,让 codex 能查文档;用户传 -m gpt-5.1-codex-max / -m gpt-5.2 时透传。

输出处理

Codex 的输出原样呈现,不允许截断、汇总、修辞——必须用 CODEX SAYS 块完整放出来。Claude 自己的解读只能附在后面,不能替代。token 数从 stderr 里 parse 出来显示成 Tokens: N。

错误处理(按作者原文)

  • 二进制找不到:Step 0 检测后停下来,给安装指引。
  • 认证错误:让用户去终端跑 codex login 走 ChatGPT 认证。
  • 外层 5 分钟超时:提示「prompt 太大或 API 慢,重试或缩范围」。
  • 内层 timeout 600 触发(exit 124):自动写 telemetry + operational learning,提示去看 ~/.codex/logs/。
  • 空响应 / session resume 失败:清掉 session 文件重来。
  • rabbit hole 检测:output 里出现 gstack-config / gstack-update-check / SKILL.md 等 gstack 文件名——意味着 codex 跑偏去读 gstack 自己的 skill 文件了,附加 warning 让用户重试。

重要规则

  • 绝不修改文件——这个 skill 是 read-only,codex 也跑在只读 sandbox。
  • 不重复 review——如果用户已经跑过 /review,这里只补独立第二意见,不要再跑 Claude 自己的 review。

适合谁

  • 想用「双 LLM 互审」减少单家 LLM 偏见的工程师
  • 重要 PR 想找一个独立大脑过一遍的代码作者
  • 难解决的 bug 想换个模型的视角

不适合

  • 没装 Codex CLI 或没认证——会直接停在 Step 0
  • 简单改动,open 一次 codex 的延迟不划算

配套

review(Claude 自审)→ codex(OpenAI 第二意见)→ requesting-code-review(让人审)三段式审阅链路。

流狐档案 运行与权限由流狐静态检查推断,不代表作者声明或实际测试结果。
领域
人工智能
作者声明的智能体
未找到明确声明;不据此推断已兼容或已测试
静态检查
94 / 100 · 启发式扫描,不代表运行安全
运行环境证据
未执行;当前没有隔离运行测试证据
作者 / 版本 / 许可证
@diegosouzapw · MIT
上下文用量评级
低消耗
接入复杂程度
需简单配置
是否需要外部接口密钥
不需要
兼容的系统
未声明(默认跨平台)
运行环境要求
未检出明确要求
检测到的文件与系统行为
  • 只读
  • 允许写入 / 修改
  • 命令执行
网络行为
仅限本地
安装命令数
无(仅作为资料)

需要注意: 未声明工具权限范围;实际权限由运行环境决定。

作者说明文件(原文) SKILL.md

# Codex Skill Guide

This skill enables the use of Codex CLI for code analysis, refactoring, and automated editing tasks.

## When to Use This Skill

Use this skill when the user:
- Asks to run Codex CLI commands (`codex exec`, `codex resume`)
- References OpenAI Codex for code analysis
- Needs automated code refactoring
- Wants AI-powered code editing

## Running a Task

1. Ask the user (via `AskUserQuestion`) which model to run (`gpt-5.2-codex` or `gpt-5.2`) AND which reasoning effort to use (`xhigh`, `high`, `medium`, or `low`) in a **single prompt with two questions**.
2. Select the sandbox mode required for the task; default to `--sandbox read-only` unless edits or network access are necessary.
3. Assemble the command with the appropriate options:
   - `-m, --model <MODEL>`
   - `--config model_reasoning_effort="<xhigh|high|medium|low>"`
   - `--sandbox <read-only|workspace-write|danger-full-access>`
   - `--full-auto`
   - `-C, --cd <DIR>`
   - `--skip-git-repo-check`
4. Always use --skip-git-repo-check.
5. When continuing a previous session:
   - Use `codex exec --skip-git-repo-check resume --last` via stdin
   - Don't use any configuration flags unless explicitly requested by the user (e.g., if they specify the model or reasoning effort when requesting to resume)
   - Resume syntax: `echo "your prompt here" | codex exec --skip-git-repo-check resume --last 2>/dev/null`
   - All flags must be inserted between `exec` and `resume`
6. **IMPORTANT**: By default, append `2>/dev/null` to all `codex exec` commands to suppress thinking tokens (stderr). Only show stderr if the user explicitly requests to see thinking tokens or if debugging is needed.
7. Run the command, capture stdout/stderr (filtered as appropriate), and summarize the outcome for the user.
8. **After Codex completes**, inform the user: "You can resume this Codex session at any time by saying 'codex resume' or asking me to continue with additional analysis or changes."

### Quick Reference

| Use case | Sandbox mode | Key flags |
| --- | --- | --- |
| Read-only review or analysis | `read-only` | `--sandbox read-only 2>/dev/null` |
| Apply local edits | `workspace-write` | `--sandbox workspace-write --full-auto 2>/dev/null` |
| Permit network or broad access | `danger-full-access` | `--sandbox danger-full-access --full-auto 2>/dev/null` |
| Resume recent session | Inherited from original | `echo "prompt" \\| codex exec --skip-git-repo-check resume --last 2>/dev/null` |
| Run from another directory | Match task needs | `-C <DIR>` plus other flags `2>/dev/null` |

## Following Up

- After every `codex` command, immediately use `AskUserQuestion` to confirm next steps, collect clarifications, or decide whether to resume with `codex exec resume --last`.
- When resuming, pipe the new prompt via stdin: `echo "new prompt" | codex exec resume --last 2>/dev/null`. The resumed session automatically uses the same model, reasoning effort, and sandbox mode from the original session.
- Restate the chosen model, reasoning effort, and sandbox mode when proposing follow-up actions.

## Error Handling

- Stop and report failures whenever `codex --version` or a `codex exec` command exits non-zero; request direction before retrying.
- Before you use high-impact flags (`--full-auto`, `--sandbox danger-full-access`, `--skip-git-repo-check`) ask the user for permission using AskUserQuestion unless it was already given.
- When output includes warnings or partial results, summarize them and ask how to adjust using `AskUserQuestion`.

## Examples

### Read-only Analysis
```bash
codex exec --skip-git-repo-check \
  -m gpt-5.2-codex \
  --config model_reasoning_effort="high" \
  --sandbox read-only \
  "Analyze the code structure and identify potential improvements" \
  2>/dev/null
```

### Apply Edits
```bash
codex exec --skip-git-repo-check \
  -m gpt-5.2 \
  --config model_reasoning_effort="medium" \
  --sandbox workspace-write \
  --full-auto \
  "Refactor the authentication module to use async/await" \
  2>/dev/null
```

### Resume Session
```bash
echo "Now add error handling to the refactored code" | \
  codex exec --skip-git-repo-check resume --last 2>/dev/null
```

## Notes

- Thinking tokens are suppressed by default using `2>/dev/null`
- Always use `--skip-git-repo-check` to avoid repository validation issues
- The `--full-auto` flag allows Codex to make changes without confirmation
- Resume sessions inherit settings from the original session

讨论

基于 GitHub Discussions。登录 GitHub 即可参与讨论、点赞、订阅更新。