Why does this skill discourage 'performative agreement'?

It prioritizes technical correctness and efficiency over social comfort, ensuring that feedback is handled through verification and implementation rather than unnecessary conversational filler.

What are Verification Gates in this skill?

Verification Gates are mandatory checkpoints that require running specific commands, such as tests or builds, to provide objective evidence before Claude can claim a task is finished.

Which development environments does the verification support?

The skill includes pre-configured patterns for Bun and NPM environments, covering testing, building, linting, and TypeScript type checking.

How does the subagent review process work?

The skill uses git SHAs to compare changes between HEAD and BASE, then dispatches a specialized code-reviewer subagent to analyze the implementation against the original requirements.

Code Review & Verification

Name: Code Review & Verification
Author: secondsky

bysecondsky

•

コラボレーションツール

Enforces technical rigor and evidence-based verification gates throughout the code review and development lifecycle.

This skill optimizes the code review process for Claude by replacing social pleasantries with technical precision. It implements a structured protocol for receiving feedback, requesting systematic subagent reviews, and enforcing 'Verification Gates' that require fresh evidence before any completion claims can be made. By prioritizing technical correctness over social comfort, it ensures that tests pass, builds succeed, and requirements are met through documented command output rather than assumptions.

主な機能

01Verification Gates requiring fresh evidence for all completion claims

02Systematic subagent review requests using git SHA comparisons

03Technical feedback protocol that eliminates performative agreement

04Quick-decision tree for handling reviewer feedback and status claims

05Standardized verification commands for Bun, NPM, and Type checking

0621 GitHub stars

ユースケース

01Processing complex feedback from external reviewers with technical evidence

02Ensuring tests and builds pass before claiming a task is complete

03Requesting deep technical audits from subagents before merging pull requests

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add secondsky/claude-skills code-review

For use in Claude.ai and ChatGPT

Download Skill