跳到主要内容
知仓学习社ZHICANG

code-review-patterns

Multi-dimensional code assessment across security, quality, performance, and maintainability with confidence-gated reporting (>=80%) and Router Cont…

不碰外部(只输出文字)无严重或高危命中a5c-ai/babysitter

它会碰到什么

扫了多少2 个文本文件,2 KB
它会碰到什么不碰外部(只输出文字)
命中总数0 处
命中统计严重 0 · 高 0 · 中 0 · 低 0

这一栏是扫描器报的事实,不是结论。命中多不等于有毒(安全工具、规则库、示例脚本本来就会包含危险写法),命中少也不等于干净。它和你手上的凭据、文件、网络有什么关系,需要你自己看。

技能内容

  • Injection vulnerabilities (SQL, XSS, command)
  • Authentication and authorization patterns
  • Secrets handling (no hardcoded credentials)
  • Input validation and sanitization

Quality (weight: 25%)

  • Naming conventions and consistency
  • Code structure (SRP, cohesion, coupling)
  • Error handling completeness
  • Type safety (no any escapes)

Performance (weight: 20%)

  • Algorithmic complexity (flag O(n^2) or worse)
  • Resource and memory leaks
  • Database query efficiency (N+1)
  • Caching opportunities

Maintainability (weight: 25%)

  • Documentation (JSDoc/TSDoc for public APIs)
  • Test coverage adequacy
  • Readability (function length, nesting depth)
  • Tech debt markers (TODO, FIXME)

Confidence Gating

  • Only report issues with confidence >= 80%
  • Empty catch blocks are always critical (100% confidence)
  • Classify: critical, high, medium, low
  • Include actionable remediation for each issue

Router Contract

Every review must produce: STATUS, BLOCKING, REQUIRES_REMEDIATION, issue counts.

When to Use

  • Code review step in BUILD workflow
  • Fix review in DEBUG workflow
  • Full REVIEW workflow

Agents Used

  • code-reviewer (primary consumer)
  • silent-failure-hunter (error handling dimension)

想直接用这个技能?

本站把开放许可(MIT / Apache 等)的技能按仓库打包整理到网盘,点一下转存到你自己的网盘,不用一个个从 GitHub 拉。许可未声明的技能只给原始仓库链接,不打包。