requesting-code-review

Run diff-based security scans, baseline-aware tests, and fail-closed JSON review on changed code.

1|Updated Apr 14, 2026
One-click install
npx skills add https://github.com/eddielueng/hermes-agent-zh --skill requesting-code-review-eddielueng
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: requesting-code-review
Source: https://github.com/eddielueng/hermes-agent-zh/tree/main/skills/software-development/requesting-code-review
Command: npx skills add https://github.com/eddielueng/hermes-agent-zh --skill requesting-code-review-eddielueng

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents bad or risky changes from being committed by running security scans, baseline-aware quality gates, and an independent reviewer before code lands.

Core Features & Use Cases

  • Static security scan of your diff: Flags likely issues like hardcoded secrets, shell injection, eval/exec patterns, unsafe deserialization, and common SQL-injection string formatting.
  • Baseline-aware tests, lint, and type checks: Runs appropriate checks per detected project type and blocks only regressions versus your pre-change baseline.
  • Independent fail-closed reviewer subagent: Reviews only your diff + static scan results and returns a strict JSON verdict; unclear output fails closed.
  • Auto-fix loop: If anything fails, a separate fix agent addresses only the reported security/logic issues and re-verifies up to two cycles.

Use case example: You make a feature change touching multiple files, then run this skill to ensure you do not introduce security problems or regressions before committing and pushing.

Quick Start

Ask Hermes Agent to request a code review verification for your current staged git changes before commit.

Frequently Asked Questions about requesting-code-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a security scan only on my git diff before committing?

Pre-commit code review prevents risky changes from being committed by running diff-based static security scans, baseline-aware tests, and an independent fail-closed JSON reviewer that blocks hardcoded secrets, shell injection, and unsafe deserialization patterns before code lands.

How does baseline-aware testing work for new code changes?

Baseline-aware testing computes pre-change failures before applying new checks, running project-specific lint and type checks to block only regressions versus your baseline rather than failing on pre-existing issues in the codebase.

Can I automatically fix security issues found during a code review?

Yes, a separate fix agent addresses only reported security or logic issues and re-verifies the changes, running up to two auto-fix and reverify cycles on specific failures detected during the independent review process.

Does the independent code reviewer fail closed on unclear output?

Yes, the independent reviewer subagent reviews your diff plus static scan results and returns a strict JSON verdict, where any unclear output fails closed to prevent ambiguous code changes from passing the review gate.

What is the best way to verify multi-file changes before pushing a PR?

The best way to verify multi-file changes before a PR is to run diff-based static security checks, baseline-aware quality gates, and an independent reviewer that targets common developer workflows like committing, pushing, or preparing pull requests after subagent-driven development.