跳到正文
TakoAPI
返回技能列表

solo-review

由 fortunto2 提供

DevOps & Cloud
404下载量由 clawskills.sh 提供
0浏览量
0 个赞登录后即可为这个技能点赞
—暂无评分
你的评分

简介

Final code review and quality gate — run tests, check coverage, audit security, verify acceptance criteria.

描述

该技能的功能

Runs a 12-dimension quality gate before shipping: tests, linting, build, security scan, acceptance criteria verification, plan completion, production logs, and commit quality checks. Produces a SHIP, FIX FIRST, or BLOCK verdict backed by actual command output. Also updates CLAUDE.md and appends fix tasks to plan.md based on findings.

Combines test runs, security audits, spec verification, and production log checks into one pass with a single go/no-go verdict, so nothing falls through the cracks between separate review steps.

何时使用

  • Final pre-deploy check after completing a feature branch
  • Verifying all spec.md acceptance criteria are actually implemented
  • Scanning a codebase for hardcoded secrets or dependency vulnerabilities
  • Checking whether recent commits follow conventional format and are atomic
  • Generating a documented audit trail before handing off a project

示例工作流

输入

User asks: Is this ready to ship?

智能体
  1. 1

    Reads CLAUDE.md and spec.md to identify acceptance criteria and detect the project stack

  2. 2

    Runs test suite, linter, type checker, and build commands for the detected stack

  3. 3

    Scans for hardcoded secrets and runs a dependency vulnerability audit

  4. 4

    Verifies each acceptance criterion in spec.md against the codebase and checks production logs

  5. 5

    Generates a report with SHIP/FIX FIRST/BLOCK verdict, per-dimension status, and file:line issue references

输出

Code review report with verdict, updated spec.md checkboxes, and fix tasks appended to plan.md if issues are found

安装

clawhub install fortunto2/solo-review

运行 CLI 命令,或把这段提示词粘贴到 Claude Code、Cursor、Codex 或你在用的任意编程智能体里。

链接