Skip to content
Back to search

agent-skills-code-review-router

https://youtu.be/Qydk2wlh4YI

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today

Check it deeper on your machine

It has changed since its scan on 11 Nov 2026.

As a Claude Code plugin, install it and then ask Claude Code:

/plugin marketplace add keithmackay/mackayi/plugin install scanskillsafety@mackayiDo a deep scan of https://github.com/win4r/agent-skills-code-review-router with scanskillsafety

Or from a terminal, clone it and run it:

git clone https://github.com/keithmackay/scanskillsafety-skill ~/.claude/skills/scanskillsafetynode ~/.claude/skills/scanskillsafety/cli.cjs --deep 'https://github.com/win4r/agent-skills-code-review-router'

It checks the exact commit: every file, its packages against OSV (api.osv.dev), how it compares with this page's scan (findsafeskills.com receives the repository's address), and, if you agree, Claude reading every file with your own Claude subscription. It never runs the code.

Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/win4r--agent-skills-code-review-router/badge.svg)](/listings/win4r--agent-skills-code-review-router)

HTML

<a href="/listings/win4r--agent-skills-code-review-router"><img src="/api/listings/win4r--agent-skills-code-review-router/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

For Claude Code

Place the skill file in one of these locations:

# Project-specific (shared with team)
.claude/skills/code-review-router/SKILL.md

# Personal (all projects)
~/.claude/skills/code-review-router/SKILL.md

For Google Antigravity

Place the skill file in one of these locations:

# Workspace-specific
<workspace-root>/.agent/skills/code-review-router/SKILL.md

# Global (all workspaces)
~/.gemini/antigravity/skills/code-review-router/SKILL.md

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 272 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. ---
  2. name: code-review-router
  3. description: Intelligently routes code reviews between Gemini CLI and Codex CLI based on tech stack, complexity, and change characteristics. Use when you want an automated code review of your current changes.
  4. ---
  5. # Code Review Router
  6. Routes code reviews to the optimal CLI (Gemini or Codex) based on change characteristics.
  7. ## When NOT to Use This Skill
  8. - For non-code reviews (documentation proofreading, prose editing)
  9. - When reviewing external/third-party code you don't control
  10. - For commit message generation (use a dedicated commit skill)
  11. - When you need a specific reviewer (use that CLI directly)
  12. ## Step 0: Environment Check
  13. Verify we're in a git repository:
  14. ```bash
  15. git rev-parse --git-dir 2>/dev/null || echo "NOT_A_GIT_REPO"
  16. ```
  17. **If not a git repo:** Stop and inform the user: "This directory is not a git repository. Initialize with `git init` or navigate to a repo."
  18. ## Step 1: Prerequisites Check
  19. Verify both CLIs are available:
  20. ```bash
  21. # Check for Gemini CLI
  22. which gemini || echo "GEMINI_NOT_FOUND"
  23. # Check for Codex CLI
  24. which codex || echo "CODEX_NOT_FOUND"
  25. ```
  26. **If neither CLI is found:** Stop and inform the user they need to install at least one:
  27. - Gemini: Check Google's Gemini CLI installation docs
  28. - Codex: Check OpenAI's Codex CLI installation docs
  29. **If only one CLI is available:** Use that CLI (no routing needed).
  30. **If both are available:** Proceed with routing analysis.
  31. ## Step 2: Analyze Git Diff
  32. Run these commands to gather diff statistics:
  33. ```bash
  34. # Get diff stats (staged + unstaged)
  35. git --no-pager diff --stat HEAD 2>/dev/null || git --no-pager diff --stat
  36. # Get full diff for pattern analysis
  37. git --no-pager diff HEAD 2>/dev/null || git --no-pager diff
  38. # Count changed files
  39. git --no-pager diff --name-only HEAD 2>/dev/null | wc -l
  40. # Count total changed lines
  41. git --no-pager diff --numstat HEAD 2>/dev/null | awk '{added+=$1; removed+=$2} END {print added+removed}'
  42. ```
  43. **If no changes detected:** Report "Nothing to review - no uncommitted changes found." and stop.
  44. ## Step 3: Calculate Complexity Score
  45. Initialize `complexity_score = 0`, then add points:
  46. | Condition | Points | Detection Method |
  47. |-----------|--------|------------------|
  48. | Files changed > 10 | +2 | `git diff --name-only \| wc -l` |
  49. | Files changed > 20 | +3 | (additional, total +5) |
  50. | Lines changed > 300 | +2 | `git diff --numstat` sum |
  51. | Lines changed > 500 | +3 | (additional, total +5) |
  52. | Multiple directories touched | +1 | Count unique dirs in changed files |
  53. | Test files included | +1 | Files matching `*test*`, `*spec*` |
  54. | Config files changed | +1 | Files: `*.config.*`, `*.json`, `*.yaml`, `*.yml`, `*.toml` |
  55. | Database/schema changes | +2 | Files: `*migration*`, `*schema*`, `*.sql`, `prisma/*` |
  56. | API route changes | +2 | Files in `api/`, `routes/`, containing `endpoint`, `handler` |
  57. | Service layer changes | +2 | Files in `services/`, `*service*`, `*provider*` |
  58. ## Step 4: Detect Language & Framework
  59. Analyze file extensions and content patterns:
  60. ### Primary Language Detection
  61. ```
  62. .ts, .tsx → TypeScript
  63. .js, .jsx → JavaScript
  64. .py → Python
  65. .go → Go
  66. .rs → Rust
  67. .java → Java
  68. .rb → Ruby
  69. .php → PHP
  70. .cs → C#
  71. .swift → Swift
  72. .kt → Kotlin
  73. ```
  74. ### Framework Detection (check imports/file patterns)
  75. ```
  76. React/Next.js → "import React", "from 'react'", "next.config", pages/, app/
  77. Vue → ".vue" files, "import Vue", "from 'vue'"
  78. Angular → "angular.json", "@angular/core"
  79. Django → "django", "models.py", "views.py", "urls.py"
  80. FastAPI → "from fastapi", "FastAPI("
  81. Express → "express()", "from 'express'"
  82. NestJS → "@nestjs/", "*.module.ts", "*.controller.ts"
  83. Rails → "Gemfile" with rails, app/controllers/
  84. Spring → "springframework", "@RestController"
  85. ```
  86. ### Security-Sensitive Patterns
  87. Detect by **file path** OR **code content**:
  88. **File paths:**
  89. ```
  90. **/auth/**
  91. **/security/**
  92. **/*authentication*
  93. **/*authorization*
  94. **/middleware/auth*
  95. ```
  96. **Code patterns (in diff content):**
  97. ```
  98. password\s*=
  99. api_key\s*=
  100. secret\s*=
  101. Bearer\s+
  102. JWT
  103. \.env
  104. credentials
  105. private_key
  106. access_token
  107. ```
  108. **Config files:**
  109. ```
  110. .env*
  111. *credentials*
  112. *secrets*
  113. *.pem
  114. *.key
  115. ```
  116. ## Step 5: Apply Routing Decision Tree
  117. **Routing Priority Order** (evaluate top-to-bottom, first match wins):
  118. ### Priority 1: Pattern-Based Rules (Hard Rules)
  119. | Pattern | Route | Reason |
  120. |---------|-------|--------|
  121. | Security-sensitive files/code detected | **Codex** | Requires careful security analysis |
  122. | Files > 20 OR lines > 500 | **Codex** | Large changeset needs thorough review |
  123. | Database migrations or schema changes | **Codex** | Architectural risk |
  124. | API/service layer modifications | **Codex** | Backend architectural changes |
  125. | Changes span 3+ top-level directories | **Codex** | Multi-service impact |
  126. | Complex TypeScript (generics, type utilities) | **Codex** | Type system complexity |
  127. | Pure frontend only (jsx/tsx/vue/css/html) | **Gemini** | Simpler, visual-focused review |
  128. | Python ecosystem (py, Django, FastAPI) | **Gemini** | Strong Python support |
  129. | Documentation only (md/txt/rst) | **Gemini** | Simple text review |
  130. ### Priority 2: Complexity Score (if no pattern matched)
  131. | Score | Route | Reason |
  132. |-------|-------|--------|
  133. | ≥ 6 | **Codex** | High complexity warrants deeper analysis |
  134. | < 6 | **Gemini** | Moderate complexity, prefer speed |
  135. ### Priority 3: Default
  136. → **Gemini** (faster feedback loop for unclear cases)
  137. ## Step 6: Execute Review
  138. ### Explain Routing Decision
  139. Before executing, output:
  140. ```
  141. ## Code Review Routing
  142. **Changes detected:**
  143. - Files: [X] files changed
  144. - Lines: [Y] lines modified
  145. - Primary language: [language]
  146. - Framework: [framework or "none detected"]
  147. **Complexity score:** [N]/10
  148. - [List contributing factors]
  149. **Routing decision:** [Gemini/Codex]
  150. - Reason: [primary reason for choice]
  151. **Executing review...**
  152. ```
  153. ### CLI Commands
  154. > **Note:** Gemini receives the diff via stdin (piped), while Codex has a dedicated `review` subcommand that reads the git context directly. If debugging, check that `git diff HEAD` produces output before running Gemini.
  155. **For Gemini:**
  156. ```bash
  157. # Pipe diff to Gemini with review prompt
  158. git --no-pager diff HEAD | gemini -p "Review this code diff for: 1) Code quality issues, 2) Best practices violations, 3) Potential bugs, 4) Security concerns, 5) Performance issues. Provide specific, actionable feedback."
  159. ```
  160. **For Codex:**
  161. ```bash
  162. # Use dedicated 'review' subcommand for non-interactive code review
  163. # Note: --uncommitted and [PROMPT] are mutually exclusive
  164. codex review --uncommitted
  165. ```
  166. ## Step 7: Handle Failures with Fallback
  167. If the chosen CLI fails (non-zero exit or error output):
  168. 1. **Report the failure:**
  169. ```
  170. [Primary CLI] failed: [error message]
  171. Attempting fallback to [other CLI]...
  172. ```
  173. 2. **Try the alternative CLI**
  174. 3. **If fallback also fails:**
  175. ```
  176. Both review CLIs failed.
  177. - Gemini error: [error]
  178. - Codex error: [error]
  179. Please check CLI installations and try manually.
  180. ```
  181. ## Step 8: Format Output
  182. Present the review results clearly:
  183. ```
  184. ## Code Review Results
  185. **Reviewed by:** [Gemini/Codex]
  186. **Routing:** [brief reason]
  187. ---
  188. [CLI output here]
  189. ---
  190. **Review complete.** [X files, Y lines analyzed]
  191. ```
  192. ## Quick Reference
  193. | Change Type | Route | Reason |
  194. |-------------|-------|--------|
  195. | React component styling | Gemini | Pure frontend |
  196. | Django view update | Gemini | Python ecosystem |
  197. | Single bug fix < 50 lines | Gemini | Simple change |
  198. | New API endpoint + tests | Codex | Architectural |
  199. | Auth system changes | Codex | Security-sensitive |
  200. | Database migration | Codex | Schema change |
  201. | Multi-service refactor | Codex | High complexity |
  202. | TypeScript type overhaul | Codex | Complex types |
README.md · README · 327 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. # [🚀观看我的哔哩哔哩视频](https://space.bilibili.com/3493277319825652)
  2. [English](#english) | [中文](#中文)
  3. ---
  4. <a name="english"></a>
  5. ### 🚀项目所对应的视频:
  6. https://youtu.be/Qydk2wlh4YI
  7. https://www.bilibili.com/video/BV1KErQB4Esx/
  8. # Code Review Router
  9. [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
  10. An **Agent Skill** that intelligently routes code reviews between Gemini CLI and Codex CLI based on tech stack, complexity, and change characteristics.
  11. ## What is an Agent Skill?
  12. Agent Skills are reusable packages of knowledge that extend what AI agents can do. A Skill is a markdown file (SKILL.md) containing instructions that the agent follows when working on specific tasks. Skills follow a **progressive disclosure** pattern:
  13. 1. **Discovery**: The agent sees available Skills with their names and descriptions
  14. 2. **Activation**: When your task matches a Skill's description, the agent loads the full instructions
  15. 3. **Execution**: The agent follows the Skill's instructions automatically
  16. This Skill works with both **Claude Code** and **Google Antigravity**.
  17. ## Why Use a Decision Tree Skill?
  18. ### Advantages of Decision Tree-Based Routing
  19. | Advantage | Description |
  20. |-----------|-------------|
  21. | **Deterministic & Predictable** | Unlike free-form prompts, decision trees follow explicit rules, ensuring consistent routing every time |
  22. | **Transparent Logic** | Every routing decision can be traced back to specific criteria—no black-box behavior |
  23. | **Optimized Tool Selection** | Automatically selects the best tool (Gemini vs Codex) based on change characteristics |
  24. | **Reduced Token Usage** | Pre-defined decision paths eliminate the need for lengthy reasoning chains |
  25. | **Extensible** | Easy to add new routing rules or modify existing ones without rewriting the entire logic |
  26. | **Fallback Handling** | Built-in graceful degradation when the primary tool fails |
  27. ### When Decision Trees Excel
  28. - **Multi-tool orchestration**: Choosing between different CLIs, APIs, or services
  29. - **Complex conditional logic**: When routing depends on multiple factors
  30. - **Audit requirements**: When you need to explain why a specific tool was selected
  31. - **Performance optimization**: Avoiding unnecessary reasoning overhead
  32. ## Features
  33. - 🔀 **Smart Routing** — Automatically selects Gemini or Codex based on code characteristics
  34. - 📊 **Complexity Scoring** — Calculates a 0-10 complexity score based on multiple factors
  35. - 🔍 **Language Detection** — Identifies TypeScript, Python, Go, Rust, and 10+ languages
  36. - 🛡️ **Security Pattern Recognition** — Detects auth, credentials, and sensitive code patterns
  37. - 🔄 **Automatic Fallback** — Switches to alternative CLI if the primary one fails
  38. - ⏱️ **Timeout Protection** — 120-second safeguard prevents hanging reviews
  39. ## Installation
  40. ### For Claude Code
  41. Place the skill file in one of these locations:
  42. ```bash
  43. # Project-specific (shared with team)
  44. .claude/skills/code-review-router/SKILL.md
  45. # Personal (all projects)
  46. ~/.claude/skills/code-review-router/SKILL.md
  47. ```
  48. ### For Google Antigravity
  49. Place the skill file in one of these locations:
  50. ```bash
  51. # Workspace-specific
  52. <workspace-root>/.agent/skills/code-review-router/SKILL.md
  53. # Global (all workspaces)
  54. ~/.gemini/antigravity/skills/code-review-router/SKILL.md
  55. ```
  56. ## Prerequisites
  57. - A Git repository with uncommitted changes
  58. - At least one CLI installed:
  59. - [Gemini CLI](https://github.com/google-gemini/gemini-cli) — Google's command-line AI assistant
  60. - [Codex CLI](https://github.com/openai/codex) — OpenAI's code review tool
  61. ## Usage
  62. Simply ask the agent to review your code:
  63. ```
  64. Review my current changes
  65. ```
  66. or
  67. ```
  68. /code-review-router
  69. ```
  70. The Skill automatically:
  71. 1. Verifies Git repository and CLI prerequisites
  72. 2. Analyzes your diff (files, lines, patterns)
  73. 3. Calculates complexity score
  74. 4. Routes to the optimal CLI
  75. 5. Presents formatted review results
  76. ## Routing Decision Tree
  77. The routing follows a priority-based decision tree:
  78. ### Priority 1: Pattern-Based Rules (Hard Rules)
  79. | Pattern | Routes To | Reason |
  80. |---------|-----------|--------|
  81. | Security-sensitive files/code | Codex | Requires careful security analysis |
  82. | Files > 20 OR lines > 500 | Codex | Large changeset needs thorough review |
  83. | Database migrations/schema | Codex | Architectural risk |
  84. | API/service layer changes | Codex | Backend architectural changes |
  85. | Changes span 3+ directories | Codex | Multi-service impact |
  86. | Complex TypeScript generics | Codex | Type system complexity |
  87. | Pure frontend (jsx/vue/css) | Gemini | Simpler, visual-focused review |
  88. | Python ecosystem | Gemini | Strong Python support |
  89. | Documentation only | Gemini | Simple text review |
  90. ### Priority 2: Complexity Score
  91. | Score | Routes To | Reason |
  92. |-------|-----------|--------|
  93. | ≥ 6 | Codex | High complexity warrants deeper analysis |
  94. | < 6 | Gemini | Moderate complexity, prefer speed |
  95. ### Priority 3: Default
  96. Falls back to **Gemini** for faster feedback on unclear cases.
  97. ## Complexity Score Calculation
  98. | Condition | Points |
  99. |-----------|--------|
  100. | Files changed > 10 | +2 |
  101. | Files changed > 20 | +3 (additional) |
  102. | Lines changed > 300 | +2 |
  103. | Lines changed > 500 | +3 (additional) |
  104. | Multiple directories | +1 |
  105. | Test files included | +1 |
  106. | Config files changed | +1 |
  107. | Database/schema changes | +2 |
  108. | API route changes | +2 |
  109. | Service layer changes | +2 |
  110. ## Quick Reference
  111. | Change Type | Route | Reason |
  112. |-------------|-------|--------|
  113. | React component styling | Gemini | Pure frontend |
  114. | Django view update | Gemini | Python ecosystem |
  115. | Single bug fix < 50 lines | Gemini | Simple change |
  116. | New API endpoint + tests | Codex | Architectural |
  117. | Auth system changes | Codex | Security-sensitive |
  118. | Database migration | Codex | Schema change |
  119. | Multi-service refactor | Codex | High complexity |
  120. ## License
  121. MIT
  122. ---
  123. <a name="中文"></a>
  124. # 代码审查路由器
  125. [![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
  126. 一个 **Agent Skill**,根据技术栈、复杂度和代码变更特征,智能地在 Gemini CLI 和 Codex CLI 之间路由代码审查。
  127. ## 什么是 Agent Skill?
  128. Agent Skills 是可复用的知识包,用于扩展 AI 代理的能力。Skill 是一个包含指令的 markdown 文件(SKILL.md),代理在处理特定任务时会遵循这些指令。Skills 采用**渐进式披露**模式:
  129. 1. **发现阶段**:代理看到可用 Skills 的名称和描述
  130. 2. **激活阶段**:当你的任务匹配某个 Skill 的描述时,代理加载完整指令
  131. 3. **执行阶段**:代理自动遵循 Skill 的指令执行任务
  132. 此 Skill 同时支持 **Claude Code** 和 **Google Antigravity**。
  133. ## 为什么使用决策树 Skill?
  134. ### 基于决策树路由的优势
  135. | 优势 | 描述 |
  136. |------|------|
  137. | **确定性与可预测性** | 与自由形式的提示不同,决策树遵循明确规则,确保每次路由结果一致 |
  138. | **逻辑透明** | 每个路由决策都可以追溯到具体条件——没有黑盒行为 |
  139. | **优化工具选择** | 根据代码变更特征自动选择最佳工具(Gemini vs Codex) |
  140. | **减少 Token 消耗** | 预定义的决策路径消除了冗长的推理链需求 |
  141. | **易于扩展** | 可以轻松添加新的路由规则或修改现有规则,无需重写整个逻辑 |
  142. | **故障回退处理** | 当主要工具失败时,内置优雅降级机制 |
  143. ### 决策树的最佳使用场景
  144. - **多工具编排**:在不同的 CLI、API 或服务之间进行选择
  145. - **复杂条件逻辑**:当路由依赖多个因素时
  146. - **审计需求**:当需要解释为什么选择特定工具时
  147. - **性能优化**:避免不必要的推理开销
  148. ## 功能特性
  149. - 🔀 **智能路由** — 根据代码特征自动选择 Gemini 或 Codex
  150. - 📊 **复杂度评分** — 基于多个因素计算 0-10 的复杂度分数
  151. - 🔍 **语言检测** — 识别 TypeScript、Python、Go、Rust 等 10+ 种语言
  152. - 🛡️ **安全模式识别** — 检测认证、凭证和敏感代码模式
  153. - 🔄 **自动故障回退** — 主 CLI 失败时自动切换到备用 CLI
  154. - ⏱️ **超时保护** — 120 秒安全机制防止审查挂起
  155. ## 安装
  156. ### Claude Code
  157. 将 skill 文件放置在以下位置之一:
  158. ```bash
  159. # 项目级(与团队共享)
  160. .claude/skills/code-review-router/SKILL.md
  161. # 个人级(所有项目)
  162. ~/.claude/skills/code-review-router/SKILL.md
  163. ```
  164. ### Google Antigravity
  165. 将 skill 文件放置在以下位置之一:
  166. ```bash
  167. # 工作区级
  168. <workspace-root>/.agent/skills/code-review-router/SKILL.md
  169. # 全局级(所有工作区)
  170. ~/.gemini/antigravity/skills/code-review-router/SKILL.md
  171. ```
  172. ## 前置条件
  173. - 包含未提交更改的 Git 仓库
  174. - 至少安装以下 CLI 之一:
  175. - [Gemini CLI](https://github.com/google-gemini/gemini-cli) — Google 的命令行 AI 助手
  176. - [Codex CLI](https://github.com/openai/codex) — OpenAI 的代码审查工具
  177. ## 使用方法
  178. 只需让代理审查你的代码:
  179. ```
  180. 审查我当前的更改
  181. ```
  182. 或者
  183. ```
  184. /code-review-router
  185. ```
  186. Skill 会自动:
  187. 1. 验证 Git 仓库和 CLI 前置条件
  188. 2. 分析你的 diff(文件、行数、模式)
  189. 3. 计算复杂度分数
  190. 4. 路由到最优的 CLI
  191. 5. 呈现格式化的审查结果
  192. ## 路由决策树
  193. 路由遵循基于优先级的决策树:
  194. ### 优先级 1:模式匹配规则(硬规则)
  195. | 模式 | 路由至 | 原因 |
  196. |------|--------|------|
  197. | 安全敏感文件/代码 | Codex | 需要仔细的安全分析 |
  198. | 文件 > 20 或行数 > 500 | Codex | 大型变更集需要彻底审查 |
  199. | 数据库迁移/Schema | Codex | 架构风险 |
  200. | API/服务层变更 | Codex | 后端架构变更 |
  201. | 变更跨越 3+ 目录 | Codex | 多服务影响 |
  202. | 复杂 TypeScript 泛型 | Codex | 类型系统复杂度 |
  203. | 纯前端 (jsx/vue/css) | Gemini | 更简单、专注视觉的审查 |
  204. | Python 生态系统 | Gemini | 强大的 Python 支持 |
  205. | 仅文档 | Gemini | 简单文本审查 |
  206. ### 优先级 2:复杂度分数
  207. | 分数 | 路由至 | 原因 |
  208. |------|--------|------|
  209. | ≥ 6 | Codex | 高复杂度需要更深入的分析 |
  210. | < 6 | Gemini | 中等复杂度,优先速度 |
  211. ### 优先级 3:默认
  212. 对于不明确的情况,回退到 **Gemini** 以获得更快的反馈。
  213. ## 复杂度分数计算
  214. | 条件 | 分数 |
  215. |------|------|
  216. | 变更文件 > 10 | +2 |
  217. | 变更文件 > 20 | +3(额外) |
  218. | 变更行数 > 300 | +2 |
  219. | 变更行数 > 500 | +3(额外) |
  220. | 涉及多个目录 | +1 |
  221. | 包含测试文件 | +1 |
  222. | 配置文件变更 | +1 |
  223. | 数据库/Schema 变更 | +2 |
  224. | API 路由变更 | +2 |
  225. | 服务层变更 | +2 |
  226. ## 快速参考
  227. | 变更类型 | 路由 | 原因 |
  228. |----------|------|------|
  229. | React 组件样式 | Gemini | 纯前端 |
  230. | Django 视图更新 | Gemini | Python 生态系统 |
  231. | 单个 bug 修复 < 50 行 | Gemini | 简单变更 |
  232. | 新 API 端点 + 测试 | Codex | 架构性 |
  233. | 认证系统变更 | Codex | 安全敏感 |
  234. | 数据库迁移 | Codex | Schema 变更 |
  235. | 多服务重构 | Codex | 高复杂度 |
  236. ## 许可证
  237. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.