Skip to content
Agentmart
Submit a tool

request-review

发起审查请求前的前置门槛与轮次规则。六条 BLOCKED 门槛不满足则请求发不出去;同型 finding 二犯触发 failure-mode audit;同对象 ≥3 轮停手升级到需求/方案层。Use when a code change has passed self-check and you are…

Skill by JianHun101

IndexedMIT
Write a review

Chalk Labs

Need help deploying request-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/jianhun101-request-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

发起审查请求前的前置门槛与轮次规则。六条 BLOCKED 门槛不满足则请求发不出去;同型 finding 二犯触发 failure-mode audit;同对象 ≥3 轮停手升级到需求/方案层。Use when a code change has passed self-check and you are about to request review. Not for writing the review conclusion, or when you are the reviewer. Output a dispatch-ready review request (template + evidence) or a BLOCKED list.

Use cases: security, writing

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/request-review

What's inside

Skill name
request-review
Description
发起审查请求前的前置门槛与轮次规则。六条 BLOCKED 门槛不满足则请求发不出去;同型 finding 二犯触发 failure-mode audit;同对象 ≥3 轮停手升级到需求/方案层。Use when a code change has passed self-check and you are about to request review. Not for writing the review conclusion, or when you are the reviewer. Output a dispatch-ready review request (template + evidence) or a BLOCKED list.
License
MIT
Instructions
101 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# request-review

发起审查请求前的门。`quality-gate` 是**自查**门(自己过),本技能是**发起侧**门(组装成可审查的形态 + 轮次裁决);两门都过,作者才自行发起。

## 何时使用

- `quality-gate` 已通过、代码已提交,准备发起审查请求
- 收到「建议修改」后返工完成,准备再次发起(轮次规则在此生效)
- 审查者第二次提出同型 finding,需要做 failure-mode audit

## 不使用的情况

- 代码还没过 `quality-gate`(先回去过门)
- 正在处理已收到的审查反馈(用 `receive-review`)
- 你自己就是审查者

## 第 1 块:BLOCKED 前置门槛(六条)

**任一条不满足 → 请求发不出去**,先补齐再发。本技能是**作者自守门**(机械阻断在入口契约层,不在技能层做)。

| #   | 门槛                  | 不满足时的处理                                                                 |
| --- | --------------------- | ------------------------------------------------------------------------------ |
| 1   | **quality-gate 通过** | 回 `quality-gate` 跑完整步骤,拿到 PASS 才算                                   |
| 2   | **测试全绿**          | `pnpm test` + `pnpm lint` 全绿;有红先修,不接受「另一条不急」                 |
| 3   | **原始需求可引用**    | 指明发起本改动的需求 / spec / 派活单(说不清要干嘛 = 没过门)                  |
| 4   | **ownership 声明**    | 写清 Cell(shared/server/web/scripts)+ Map Delta(新增/删除/移动的文件)      |
| 5   | **前端真机自证**      | 动了前端 → 启动应用走通至少一条真实用户路径,附路径与结果                      |
| 6   | **根目录工件闸门**    | 临时文件 / 调试产物不入库:无未跟踪垃圾、`.gitignore` 覆盖到位、提交限定了路径 |

## 第 2 块:R2+ 同型 finding → 强制 failure-mode audit

同一类问题被**第二次**判出来(R2+)→ **不许单点修**。

1. **先扫全 diff**:把这一型在整份 diff 里找全,列成清单(不止被点到的那一处)
2. **问「为什么会产生」**:找出生成这只问题的模式,判断它是孤例还是机制
3. **一次修一族**:清单里每一处同批修掉;修不掉的作为 Open Question 点名具体文件/符号
4. **回复里报清单**:报「同型共 N 处 / 修了 M 处 / 余 K 处理由」

**判据**:同型问题第二次出现 = 上一轮是**点修**而不是**族修**——这是方法的缺陷,不是手滑。

## 第 3 问:这次撞出的东西要不要上浮?

发起前自问一次。**非门槛**——有就落,没有就过,不阻塞发起。

| 撞出的东西                                         | 落点                                            |
| -------------------------------------------------- | ----------------------------------------------- |
| **否决理由** —— 考虑过又放弃的方案、被推翻的旧口径 | `docs/adr/`(过 ADR 准入门槛才落)              |
| **踩坑教训** —— 撞出来的、当时没有备选方案可比的   | `docs/lessons/`(内容边界见该目录 `README.md`) |
| **操作规则** —— 以后每次都得这么做 / 不能这么做    | `AGENTS.md` Conventions                         |

本问只做**指向**——规则正文归落点所在文件,此处不重述。

## 第 4 块:同对象 ≥3 轮 → 停手升级

同一个对象(同一 commit / 同一文件 / 同一功能点)**第 3 轮**被打回 → **停止继续修**,升级到需求·方案层。

- 同一对象上打转三轮 = 问题不在实现,在**需求没钉死**或**方案选错了**
- 动作:把「三轮各修了什么、每轮为什么还不够」列出来,回到需求 / spec / 方案层重新对齐,再决定是继续修还是重做

**F229 教训**:同一对象 20 轮返工,根因是最初的需求分歧从未被摆到台面上。

## 第 5 块:模板

审查请求的正文用 `skills/refs/review-request-template.md`(共享模板,本技能**不重写**):What / Why / Original Req / Tradeoff / Architecture Ownership / Open Questions / Reviewer Checklist / Self-Check Evidence。

- 模板里的 `{}` 占位符**逐项填实**,不留空壳
- Self-Check Evidence 必须是**跑过的真结果**(测试数、lint 输出、真机路径),不是「应该是通过」

## 输出

```
## Review Request Readiness

### 门槛(六条)
✅ 全过 / ❌ 未过:<第 N 条 + 缺什么>

### 轮次
第 N 轮 · 同型 finding:<无 / 有 → 已做 failure-mode audit:同型 N 处,修 M,余 K>

### 请求正文
<按 skills/refs/review-request-template.md 填好的正文>
```

## Common Mistakes

| 错误                                   | 正确做法                                    |
| -------------------------------------- | ------------------------------------------- |
| 测试还有红的就发请求(「另一条不急」) | 六条门槛全过才发;不过就是 BLOCKED          |
| 被指出同型问题第二次,只修点到的那处   | 先扫全 diff 列同型清单,一次修一族          |
| 第 3 轮还在原地改细节                  | 停手,升级到需求 / 方案层重新对齐           |
| 请求里写「应该没问题」                 | 附跑过的真结果:测试数、lint 输出、真机路径 |
| 自己去写审查结论                       | 结论归审查者;本技能只管发起侧              |

## 与其他 skill 区别

| skill          | 区别                                                                          |
| -------------- | ----------------------------------------------------------------------------- |
| quality-gate   | quality-gate 是自查门(自己过);request-review 是发起侧门(组装 + 轮次裁决) |
| receive-review | request-review 管发出去之前;receive-review 管收到反馈之后                    |

From JianHun101/catStudio, MIT.

More from JianHun101/catStudio

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
Sep 21, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

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

More from JianHun101

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 request-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 request-review?

CallGet help
Cite this page

request-review skill. Agent Mart. https://agentmart.dev/skills/jianhun101-request-review (last updated 2026-10-24)