review-changes

Reviews code changes using risk-scored change detection and knowledge graph impact analysis.

25|4|Updated Dec 21, 2025
One-click install
npx skills add https://github.com/thangchung/agent-engineering-experiment --skill review-changes-thangchung
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: review-changes
Source: https://github.com/thangchung/agent-engineering-experiment/tree/main/agentgateway-entraid-obo/.claude/skills/review-changes
Command: npx skills add https://github.com/thangchung/agent-engineering-experiment --skill review-changes-thangchung

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Manual code reviews often miss hidden dependencies and untested changes, leading to regressions in production. This Skill performs a structured, risk-aware review by detecting changes, tracing impacted execution paths, and checking test coverage before merge. ## Core Features & Use Cases - Risk-Scored Change Detection: Runs change detection to classify modifications by risk level (high/medium/low). - Impact Analysis: Identifies affected execution flows and calculates the blast radius of each change using a knowledge graph. - Test Coverage Verification: Checks whether high-risk functions have corresponding tests and suggests specific test cases for gaps. - Use Case: Before merging a pull request that modifies authentication logic, run this Skill to see which downstream flows are affected, whether tests exist for the changed functions, and receive a merge recommendation grouped by risk level. ## Quick Start Review the current uncommitted changes and give me a risk-grouped report with test coverage status and a merge recommendation.

Frequently Asked Questions about review-changes

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

FAQPage Schema
How do I review code changes for risk before merging?▼

Run change detection to get a risk-scored analysis of modifications, then trace affected execution flows and check test coverage for high-risk functions. The review output groups findings by risk level with a merge recommendation.

How to find which code paths are affected by a change?▼

Use the affected flows analysis to identify execution paths impacted by modified functions, then check the impact radius to understand the full blast radius. This reveals downstream callers that may break.

Can this check test coverage for changed functions?▼

Yes, it queries the knowledge graph for tests associated with each high-risk function. For untested changes, it suggests specific test cases to close coverage gaps before merging.

What are the limitations of graph-based code review?▼

Graph-based review depends on an up-to-date knowledge graph of the codebase, so stale or incomplete graphs may miss impacts. It also cannot judge business logic correctness, only structural risk and coverage.

How does the review stay within token limits?▼

It starts with minimal context retrieval and uses the minimal detail level on all graph calls, escalating only when needed. The target is completing a review in five or fewer tool calls and under 800 output tokens.