Token导航 LogoToken导航TokenDH.com
研究检索只读github未标认证来源可访问许可证需确认审计通过

review审查

Agent Skill

review 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

220

周安装

9

GitHub Stars

2

下载量

71
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:review(审查)
来源仓库:https://github.com/cniska/skills
仓库路径:skills/review
安装命令:
npx skills add https://github.com/cniska/skills --skill review
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/cniska/skills --skill review

简介

review 对当前分支变更进行全面质量审查,覆盖多个维度但避免重复发现问题。

  • 适用于 ~100~300 行逻辑变更的快速评审,超千行建议拆分为小提交再审。
  • 仅审查相对于 main 的差异,但会读取周边代码理解上下文与边界。
  • 批准标准是整体提升代码健康度,允许局部不完美以换取进度推进。
  • review 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Review

Run all review dimensions against the current branch and produce one unified review. Approve when a change improves overall code health, even if it isn't perfect.

Scope

Review only what changed on the current branch against main, but read enough surrounding code and docs to understand conventions and boundaries.

Do not duplicate the same issue across categories.

Change sizing

Before reviewing, check the diff size:

  • ~100 lines: good, reviewable in one pass.
  • ~300 lines: acceptable if one logical change.
  • ~1000 lines: too large — ask the author to split before reviewing.

Refactoring mixed with feature work is two changes. Flag it.

Workflow

  1. Determine diff scope: git log main..HEAD --oneline and git diff main...HEAD --stat.
  2. If no commits ahead of main, report that and stop.
  3. Read changed files in full, plus any project-level convention docs.
  4. Review tests first — they reveal intent and coverage gaps.
  5. Run each review dimension: Style, Architecture, Docs, Security, Tests. Load the corresponding skill for each dimension when possible.
  6. Merge findings: deduplicate, keep strongest framing per root issue.
  7. Label every finding by severity:
LabelMeaningAction
*(unmarked)*Must fixAddress before merge
Critical:Blocks mergeSecurity, data loss, broken functionality
Nit:OptionalStyle preference, minor improvement
Consider:SuggestionWorth thinking about, not required
  1. Order by merge relevance: must-fix → should-fix → optional.

Dependency review

If the change adds a dependency, check:

  • Does the existing stack already solve this?
  • Is it actively maintained?
  • What's the size impact?
  • Any known vulnerabilities?

Every dependency is a liability.

Fix policy

Default to fixing all findings — including trivial ones. Small issues left unfixed accumulate into tech debt. If a finding is worth reporting, it's worth fixing before merge.

Output

One section per review dimension (Style, Architecture, Documentation, Security, Tests), then a summary table: category | must-fix | should-fix | optional. Note categories with no findings.

See also

  • style, architecture, docs, security, tests for dimension-specific depth
  • references/security-checklist.md for concrete abuse-path checks
  • references/testing-patterns.md for test-quality review heuristics

Red flags

  • Reviewing only the diff without reading touched files in context
  • Duplicating the same root issue across categories
  • Generic cleanup wishlists
  • Speculative issues without evidence
  • Broad rewrite suggestions out of scope
  • "LGTM" without evidence of review
  • Softening real issues — if it's a bug, say so directly
  • Accepting "I'll fix it later" — require cleanup before merge

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.83%
按下载量换算24

Claude

32.14%
按下载量换算23

Cursor

21.66%
按下载量换算15

Gemini CLI

9.39%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills