Skip to content
Agentmart
Submit a tool

generic-review

通用代码审查 — Claude 内部三路并行(code + security + arch)+ 可选第四路跨厂商对抗审查(codex cli),汇总对比。

Skill by Brothelmdzz

IndexedMIT★ 2
Write a review

Chalk Labs

Need help deploying generic-review?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/brothelmdzz-generic-review

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

通用代码审查 — Claude 内部三路并行(code + security + arch)+ 可选第四路跨厂商对抗审查(codex cli),汇总对比。

Use cases: security, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/generic-review

What's inside

Skill name
generic-review
Description
通用代码审查 — Claude 内部三路并行(code + security + arch)+ 可选第四路跨厂商对抗审查(codex cli),汇总对比。
License
MIT
Instructions
129 lines in SKILL.md · view on GitHub

Agent Skills spec issues

  • files: referenced but missing: scripts/detect-codex.sh

Files (3)

  • SKILL.md
  • playbook-fallback.md
  • report-template.md

SKILL.md (first lines)

# 通用代码审查

## 约束

- **Layer 1 必跑**:Claude 三路并行(code + security + arch),不得串行
- **Layer 2 可选**:检测到 codex cli 且配置允许时启用跨厂商审查
- 每路独立输出报告,汇总时保留多方对比
- CRITICAL 问题必须自动修复后重编译验证
- 报告文件必须非空(>500 字节 + 至少 2 个 ## 标题),否则 stop-hook 会打回

## Layer 1: Claude 内部三路并行(必跑)

用 `Agent(run_in_background=true)` 同时启动:

| Agent | 模型 | 侧重 | 产出 |
|-------|------|------|------|
| code-reviewer | sonnet | 代码质量、bug、边界条件、可维护性 | `.claude/reports/review-code.md` |
| security-reviewer | sonnet | SQL 注入、XSS、敏感信息泄露、权限缺失 | `.claude/reports/review-security.md` |
| architect | opus | 架构合理性、模块边界、接口设计 | `.claude/reports/review-arch.md` |

模型异构(sonnet × 2 + opus × 1)= 消除单模型偏见 + 成本反而比全 opus 低。

每个问题标注 CRITICAL / WARN / INFO 级别。

## Layer 2: 跨厂商对抗审查(可选)

### 启用条件

满足**所有**:

1. `dev-config.yml` 中 `review.cross_vendor.enabled` 为 `auto`(默认)或 `true`
2. `bash ${CLAUDE_PLUGIN_ROOT}/scripts/detect-codex.sh` 退出码 0(codex cli 就绪 + 配置完整)

### 检测命令

```bash
bash "${CLAUDE_PLUGIN_ROOT}/scripts/detect-codex.sh"
# 退出码 0 = 就绪,可启用 Layer 2
# 退出码 1 = 未装或未配置
# 退出码 2 = enabled=true 但未就绪(必须报错)
```

### 调用

⚠️ 不要用 `codex review` 直接 redirect — 它的输出含 banner 和 deprecated 警告,需要后处理。
✅ 用 `codex exec --output-last-message`,直接把审查结论写入文件(实测干净 markdown):

```bash
# 在项目根执行
BASE_BRANCH=$(git rev-parse --abbrev-ref origin/HEAD 2>/dev/null | sed 's|origin/||' || echo master)
mkdir -p .claude/reports

codex exec \
  --skip-git-repo-check \
  --output-last-message .claude/reports/review-cross-vendor.md \
  -s read-only \
  "请对 git diff ${BASE_BRANCH}...HEAD 做代码审查。

## 任务
作为独立的非 Claude 模型,审查这次变更,找出:
- CRITICAL:必须修复的 bug、安全漏洞、严重逻辑错误
- WARN:建议修复
- INFO:观察/建议

## 输出格式(严格 markdown)

## CRITICAL
| 文件:行号 | 问题 |
|----------|------|

## WARN
| 文件:行号 | 问题 |
|----------|------|

## INFO
- ...

## 跨厂商视角总结
你作为非 Claude 模型,从你的训练数据 / 推理路径角度,这次变更最值得人类审查者注意的 1-2 个点是什么?" \
  || echo "cross-vendor review failed, fallback to Layer 1 only" \
     > .claude/reports/review-cross-vendor.md
```

**关键 flag 说明**:
- `--skip-git-repo-check`: 不要求项目在 codex trust list(避免每个新项目都要先 trust)
- `--output-last-message <FILE>`: 仅把最终回复写文件(不含 codex 自身 banner、deprecated 警告、tool call 日志)
- `-s read-only`: 沙箱限制,审查不修改文件
- `--config model="..."`(可选): 临时覆盖 `~/.codex/config.toml` 中的 model

产出:`.claude/reports/review-cross-vendor.md`,使用 codex 配置的非 Claude 模型(如 GPT-5 / GLM / o3 等,由用户 `~/.codex/config.toml` 中 `model_provider` + `model` 决定)。

### enabled 取值

| 值 | 检测到 codex | 行为 |
|---|-------------|------|
| `auto`(默认)| ✅ 装了 | 启用 Layer 2 |
| `auto` | ❌ 没装 | 静默跳过 |
| `true` | ❌ 没装 | 报错,要求安装 codex 或改为 false/auto |
| `false` | -- | 永远跳过 |

## 汇总规则

按以下顺序处理:

1. **去重**:按 `file:line + 问题摘要哈希` 去重
2. **置信度判定**:
   - 任意 ≥ 2 路命中同一问题 → **高置信**,必须修
   - 仅 1 路 Claude reviewer 命中 → 中置信,标注待人决
   - 仅 cross-vendor 单独命中 → **单独 section "异构视角发现"**(不强制修,但保留报告)
3. **CRITICAL 处理**:任一路标 CRITICAL 且置信度高 → 自动修复 → 重编译确认
4. **结论**:无 CRITICAL 高置信问题 → 通过

## 参考

- 报告格式 → 读取 [`report-template.md`](report-template.md)
- Agent 启动失败 / 降级 → 读取 [`playbook-fallback.md`](playbook-fallback.md)
- codex 检测脚本 → [`scripts/detect-codex.sh`](../../scripts/detect-codex.sh)

## 产出

From Brothelmdzz/dev-harness, MIT.

More from Brothelmdzz/dev-harness

Signals

Stars
2
Forks
0
Open issues
0
Last commit (this folder)
May 8, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from Brothelmdzz

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get generic-review running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying generic-review?

CallGet help
Cite this page

generic-review skill. Agent Mart. https://agentmart.dev/skills/brothelmdzz-generic-review (last updated 2026-10-24)