What are verification gates in this skill?

Verification gates are mandatory checks that require the agent to run specific commands and provide actual output evidence before claiming a task or build is successful.

How does the subagent review process work?

The skill dispatches a specialized code-reviewer subagent to analyze the diff between specific Git SHAs against the original plan or requirements to ensure nothing was missed.

Why does the skill discourage performative agreement?

To ensure technical correctness. It forces the agent to focus on technical logic and verification rather than using polite filler phrases like 'Great point' or 'You're right'.

Does this skill work with Git?

Yes, it specifically uses Git commands to identify BASE and HEAD SHAs to provide context for reviews and ensure only current changes are analyzed.

When is the best time to trigger the code-review skill?

It should be triggered when receiving feedback, after completing any major task or feature, and immediately before making any completion or success claims.

Technical Code Reviewer

Name: Technical Code Reviewer
Author: hongbietcode

byhongbietcode

•

Collaboration Tools

Enforces technical rigor and evidence-based verification during code reviews to prevent false completion claims and ensure software quality.

This skill transforms the code review process from a social interaction into a rigorous technical validation protocol. It emphasizes honest, concise feedback, systematic subagent reviews using Git diffs, and mandatory verification gates that require hard evidence before any task is marked as complete. By prioritizing technical correctness over social comfort and strictly adhering to YAGNI, KISS, and DRY principles, it ensures that pull requests and features meet high-quality standards while preventing the common pitfall of performative agreement.

Key Features

01Integration with subagent-driven development for granular, task-by-task quality control

02Anti-performative feedback protocols that prioritize technical logic over social politeness

032 GitHub stars

04Strict verification gates requiring command output evidence before any status claims

05Automated decision trees for handling human and external reviewer feedback

06Systematic subagent-driven reviews using Git SHAs for precise code diff analysis

Use Cases

01Ensuring all tests and builds pass with verifiable proof before updating project status

02Managing and prioritizing feedback from external reviewers while maintaining architectural integrity

03Validating complex refactors or new features before merging into the main branch

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add hongbietcode/synthetic-claude code-review

For use in Claude.ai and ChatGPT

Download Skill