Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/lion-1209/coderio/code-reviewnpx skills add Lion-1209/coderio --skill code-reviewgit clone --depth 1 https://github.com/Lion-1209/coderioWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00023 | $0.02274 |
| Opus 5 | $0.00012 | $0.01137 |
| Sonnet 5 | $0.00005 | $0.00455 |
| Haiku 4.5 | $0.00002 | $0.00227 |
Grade A, and why
code-review scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
Copies of this mod
1 near-identical copy found in the catalogue:
- code-review — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Review
概述
对代码做有重点、分层次的审查——先抓正确性和安全问题,再管可读性,风格交给工具。核心:好的 review 让作者知道"哪里有真问题、该先改什么";坏的 review 是一堆 nit 淹没重点,或一句"看起来不错"放过错误。
何时使用
- 提交前审查自己或别人的代码(自我 review 见下文专门技巧)
- 同事让你看他的 PR
- 不确定 review 该看什么、怎么反馈
不该用:纯风格/格式问题(交给 linter/formatter 自动化,不该占用人 review);想确认"功能对不对"——那是 verify-and-fix(实际跑)的事,review 是静态审查、替代不了运行验证。
与相邻 skill 的边界:code-review 是审查视角(找问题、分严重度、给建议),verify-and-fix 是修复视角(改对、验证)。code-review 发现的问题,需要修的进入 verify-and-fix 流程;review 中发现的"行为不对/报错"可能要先用 debugging 定位根因。三者接力:review 找问题 → debugging 定位 → verify-and-fix 修+验证。
核心内容
正确性优先,别被风格带偏
review 最常见的错是逐行挑风格(命名、缩进、注释多少)而放过正确性(逻辑对不对、边界处理了吗、有没有竞态)。正确性问题会让程序出错,风格问题只是不好看——前者 critical,后者 nit,优先级天差地别。
按重要性排序,review 该查的维度:
- 正确性:逻辑对吗?能跑通预期路径吗?有没有逻辑漏洞?(最该花时间)
- 边界与异常:空值/空集合/零/负数/超大输入怎么处理?外部依赖失败(网络/DB)呢?
- 安全:有没有注入(SQL/命令/XSS)?密码/密钥处理对吗?权限检查到位吗?敏感信息泄露吗?
- 并发:共享状态有竞态吗?异步顺序依赖对吗?资源泄漏(未关闭的连接/锁未释放)?
- 可维护性:命名清不清楚?结构是否过度复杂?有没有重复?(这里才是 nit 的领地)
- 测试:有测试吗?测的是行为还是实现?覆盖了关键路径和边界吗?
前 4 类是"会让程序出错或出事"的,必须查;第 5 类是"让人难受"的,次要;风格细节(缩进/格式)不该人查,交给工具。
按改动性质调整重点:上面是通用清单,但不同代码该重点查的不同——别对所有代码平均用力。涉及钱/库存/计数的,重点查原子性和一致性(中途失败会不会凭空产生/消失);涉及外部输入的(用户输入、API、文件),重点查安全(注入、越权);涉及共享状态/异步的,重点查竞态和资源泄漏;涉及配置/迁移的,重点查回滚和兼容。先识别"这段代码的风险面在哪",把 review 力量集中投到那里。
分严重度,别把 nit 和 critical 混着
review 反馈必须分严重度,让作者知道先改什么:
- 阻塞(blocking / critical):必须改才能合并——逻辑错、安全漏洞、会崩溃、数据丢失风险。
- 重要(important):强烈建议改——边界没处理、缺少测试、设计有隐患,但不阻塞本次合并。
- 建议(nit / suggestion):可选——命名、可读性、小重构。改了更好,不改也能过。
不分层的 review 有两种失败:把 nit 当 critical(作者被一堆小事压垮,反而漏改真问题)、把 critical 当 nit(真问题被淹没在风格意见里)。nit 要克制——堆 15 条 nit 是在浪费作者时间,把同类 nit 归并成一条"风格建议",或直接交给 linter。
反馈要带依据和建议,不只是"这不好"
每条 review 意见应该让作者能理解问题 + 知道怎么改:
- 指出问题 + 为什么是问题:不说"这写得不好",说"
user.name在 user 为 null 时会抛 TypeError——db.find 找不到时返回 null"。 - 给方向或示例:不只说"改一下",给"加个 null 检查,找不到时抛业务错误或返回 null,看调用方期望"。
- 区分事实和偏好:"这里有 null 风险"(事实,基于代码)vs "我觉得该用 early return"(偏好,标注是建议)。
带依据的反馈让作者能判断对错、学到东西;空泛的"不好"让作者只能盲从或抵触。
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 103 lines · 23 tokens per session scan A d60a60059723
code-review is a skill published in the GitHub repository Lion-1209/coderio (9 stars, last pushed 5d ago), licensed MIT. It adds 23 tokens to every session and 2,274 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
blog-writer
Peri 项目博客写作风格指南。当用户说"写博客"、"写文章"、"出稿"、 "按风格写"、"博客"时触发。也适用于用户丢过来素材说"帮我写篇博客"的场景。 覆盖项目介绍、技术复盘、架构讨论、性能优化、架构设计等类型。.
auto-devflow
Use when starting an issue, bugfix, feature, or refactor that benefits from an adaptive development workflow. Select lite, normal, pro, max, or ultra from task complexity and risk, then use only the coordination, review, and verification phases that the task actually needs.
langfuse
Interact with Langfuse and access its documentation. Use when needing to (1) query or modify Langfuse data programmatically via the CLI — traces, prompts, datasets, scores, sessions, and any other API resource, (2) look up Langfuse documentation, concepts, integration guides, or SDK usage, or (3) understand how any…
self-build
Builds isolated npm capability packages that operate on real project code and connects them to Peri through MCP/MCPP and MetaHarness. Use when adding tools, resources, remote skills or agents, creating a Bun/Node.js stdio server, linking .mcp.json, or changing the active prompt and middleware set.
auto-issue-fixer
Issue 全生命周期管理——从创建到归档。当用户描述技术问题、提 bug、"帮我记录"、 "修一下 X issue"、"验证一下"、"归档 issue"时立即触发。单入口自动分发, 替代旧 issue-create/fix-issue/issue-verify/issue-archive 四个技能。 即使用户没有用"issue"这个词,只要在描述值得追踪的技术问题就应触发。.
llm-log-analyzer
分析 llm-gateway 代理产生的请求/响应日志。当用户说"分析日志"、"查看 LLM 请求"、"对比 session"、"检查 token 用量"、"日志里有什么"、"帮我看看 data 目录"、"哪个请求失败了"、"找一下 session 的请求"等涉及 LLM 网关日志分析的场景时使用此 skill。即使用户只是笼统地说"看看日志"或"data 里有什么",也应触发。.