braydio/CUSTODIAN / .claude/skills/review-changes
review-changes
VerifiedPerform a structured code review using change detection and impact
Review Changes
Perform a thorough, risk-aware code review using the knowledge graph.
Steps
- Run
detect_changes_toolto get risk-scored change analysis. - Run
get_affected_flows_toolto find impacted execution paths. - For each high-risk function, run
query_graph_toolwith pattern="tests_for" to check test coverage. - Run
get_impact_radius_toolto understand the blast radius. - For any untested changes, suggest specific test cases.
Output Format
Provide findings grouped by risk level (high/medium/low) with:
- What changed and why it matters
- Test coverage status
- Suggested improvements
- Overall merge recommendation
Token Efficiency Rules
- In the persistent coordination checkout, start with
get_minimal_context(task="<your task>")when the graph is ready. - In a linked/ephemeral worktree, do not cold-build a graph merely because the first call reports
missing_graph,not_ready, empty/unmapped, or stale coverage. If baseline structural context is useful, point CRG at the persistent coordination root; then verify every implementation/review fact against the current worktree. - If the graph is unusable after one cheap attempt, immediately use targeted symbol/path reads. Current worktree files and
git diffare authoritative for branch-local changes. - When the task packet provides a bounded
## Context Pack, generate it once from the claimed worktree withscripts/ai/pack-context.sh task ...and use it instead of broad file dumping. - Use
detail_level="minimal"on graph calls. Escalate only when minimal is insufficient. - Target: complete any review/debug/refactor task in ≤5 graph calls and ≤800 graph-output tokens.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| receiving-code-reviewobra/superpowersUse when receiving code review feedback, before implementing suggestions, especially if feedback seems unclear or technically questionable - requires technical rigor and verification, not performative agreement or blind implementation | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| requesting-code-reviewobra/superpowersUse when completing tasks, implementing major features, or before merging to verify work meets requirements | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| using-git-worktreesobra/superpowersUse when starting feature work that needs isolation from current workspace or before executing implementation plans - ensures an isolated workspace exists via native tools or git worktree fallback | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |
| verification-before-completionobra/superpowersUse when about to claim work is complete, fixed, or passing, before committing or creating PRs - requires running verification commands and confirming output before making any success claims; evidence before assertions always | Stars: 319k | Updated | Claude CodeCodex | Verified | Install |