review-changes — Claude Code skill: what it does, install
Reviews a change set through a knowledge graph, scoring risk and tracing impact. Use it to inspect changed functions, affected flows, and missing tests.
npx skills add tirth8205/code-review-graph --skill review-changes -a claude-codeWhat review-changes does
This skill reviews a change set with a knowledge graph. It identifies changed functions with risk scores, test gaps, and affected flows, then groups findings as high, medium, or low risk with coverage, a suggested fix, and a merge recommendation.
It starts with get_minimal_context_tool, then calls get_affected_flows_tool, query_graph_tool with the tests_for pattern, or get_impact_radius_tool when needed. A budget of about five tool calls and 800 tokens of graph output means the review stays targeted. Read the implementation and its tests before making code changes.
What it helps with
- identify changed functions with high risk
- find test gaps after a code change
- check test coverage for risky functions
- trace the impact radius of a change
- suggest test cases for untested code
Example prompts for the agent
- Review the current change set with the `review-changes` skill.
- Run `review-changes` and group the findings by risk level.
- Use `review-changes` to inspect affected flows and test coverage for these changes.
How to install review-changes
1Skills CLI (Vercel)
npx skills add tirth8205/code-review-graph --skill review-changes -a claude-code2Manually: put the files in place
~/.claude/skills/review-changes/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/review-changes/
1Skills CLI (Vercel)
npx skills add tirth8205/code-review-graph --skill review-changes -a codex2Manually: put the files in place
.agents/skills/review-changes/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add tirth8205/code-review-graph --skill review-changes -a cursor2Manually: put the files in place
.cursor/skills/review-changes/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install tirth8205/code-review-graph review-changes2Manually: put the files in place
.github/skills/review-changes/SKILL.md1Skills CLI (Vercel)
npx skills add tirth8205/code-review-graph --skill review-changes -a gemini-cli2Platform CLI
gemini skills install https://github.com/tirth8205/code-review-graph --path skills/review-changesManually: put the files in place
.devin/skills/review-changes/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add tirth8205/code-review-graph --skill review-changes -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
From the same set7
The whole set →Similar
Questions about review-changes
How do I install review-changes in Claude Code?
Run npx skills add tirth8205/code-review-graph --skill review-changes -a claude-code, then run npx skills add tirth8205/code-review-graph --skill review-changes -a claude-code again. The skill file is at ~/.claude/skills/review-changes/SKILL.md.
What does review-changes do?
It performs a structured review of a change set, scoring changed functions and checking test gaps and affected flows. Its output ends with a merge recommendation.
Why does review-changes still require reading the code and tests?
The knowledge graph narrows the review scope, but it does not replace the implementation or its tests. The skill instructs the agent to read both before changing code.
Source text
SKILL.md · MIT
--- name: review-changes description: Perform a structured code review using change detection and impact --- ## Review Changes Review a change set with risk scores and blast radius from the knowledge graph. ### Steps 1. Call `detect_changes_tool` for risk-scored changed functions, test gaps and affected flows. 2. Call `get_affected_flows_tool` only when you need the steps of an affected flow. 3. For each high-risk function, call `query_graph_tool` with `pattern="tests_for"` to check test coverage. 4. Call `get_impact_radius_tool` when the blast radius is not clear from step 1. 5. Suggest specific test cases for untested changes. ### Output Format Group findings by risk level (high, medium, low). For each finding give what changed and why it matters, its test coverage, and the suggested fix. End with a merge recommendation. ## Token Efficiency Rules - Call `get_minimal_context_tool(task="<your task>")` before any other graph tool. - Pass `detail_level="minimal"` wherever a tool accepts it. Use "standard" only when minimal is not enough. - Prefer a targeted `query_graph_tool` call over a broad listing call. - Budget: about five tool calls and 800 tokens of graph output per task. - Read the implementation and its tests before changing code. The graph narrows scope; it does not replace the source.
© tirth8205 · MIT · tirth8205/code-review-graph