Can it check for missing tests?

Yes, the skill queries the graph to see if modified functions have corresponding tests and flags any untested changes as a high-risk finding.

Do I need to manually update the code graph?

No, the skill includes a tool to incrementally update the local knowledge graph automatically based on your latest git changes before the review begins.

What is 'blast radius' detection?

Blast radius detection identifies functions, classes, or files that might break because they depend on the code you just changed, ensuring you don't miss unintended side effects.

Does this work with any programming language?

As long as the underlying knowledge graph tool supports the language's structure and dependency mapping, Review Delta can analyze it.

How does Review Delta save tokens?

Instead of sending your entire codebase to the AI, it uses a local knowledge graph to identify only the changed code and its immediate 2-hop dependencies, reducing token usage by up to 49x.

Impact-Aware Code Review (Review Delta)

Name: Impact-Aware Code Review (Review Delta)
Author: tirth8205

bytirth8205

•

10,024

•

协作工具

Performs focused, token-efficient code reviews by analyzing git diffs and their downstream blast radius via a local knowledge graph.

Review Delta is a specialized Claude Code skill designed to optimize the code review process by focusing only on what matters. By building a persistent local knowledge graph of your codebase, it identifies the "blast radius" of any change—mapping out impacted callers, inheritance chains, and dependent files. This allows Claude to provide deep structural insights while using up to 49x fewer tokens than traditional full-context reviews. It is particularly effective for identifying regression risks in complex systems and ensuring that new changes are adequately covered by tests without needing to upload the entire repository context.

主要功能

01Local knowledge graph for persistent codebase mapping

02Structured risk assessment (Low/Medium/High) for every review

03Automated blast-radius detection for downstream dependencies

0410,024 GitHub stars

05Token-optimized delta reviews via git diff analysis

06Integrated test coverage verification for modified functions

使用场景

01Identifying side effects in downstream callers after a major function refactor

02Verifying test coverage specifically for newly added or modified logic

03Conducting efficient reviews on large pull requests without exceeding context limits

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tirth8205/code-review-graph review-delta

For use in Claude.ai and ChatGPT

主要功能

01Local knowledge graph for persistent codebase mapping

02Structured risk assessment (Low/Medium/High) for every review

03Automated blast-radius detection for downstream dependencies

0410,024 GitHub stars

05Token-optimized delta reviews via git diff analysis

06Integrated test coverage verification for modified functions

使用场景

01Identifying side effects in downstream callers after a major function refactor

02Verifying test coverage specifically for newly added or modified logic

03Conducting efficient reviews on large pull requests without exceeding context limits

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tirth8205/code-review-graph review-delta

For use in Claude.ai and ChatGPT