What problem does it solve? Code reviews often miss breakage that grep and symbol searches cannot reveal, such as changes to API responses, database columns, wire formats, or behavior three hops downstream. This Skill finds what a change could break elsewhere before it ships and proves the key safety assumption with executable evidence instead of a convincing-sounding writeup. ## Core Features & Use Cases - Hidden Breakage Detection: Traces risks beyond the diff, including library internals, pinned versions, timing behavior, feature flags, and cross-language consumers of the same data. - Evidence-Based Verification: Ranks each safety claim on a five-level confidence ladder and requires proving the critical fact with a runnable script or test that calls the real code. - Structured Risk Report: Returns what changed, the one fact the change is safe because of, confirmed risks with file:line citations, cleared concerns, and the cheapest pre-merge test. - Use Case: Before merging a small diff you do not trust, ask for its blast radius and receive a verified analysis where the core safety claim is either proven by executed code or explicitly marked unproven. ## Quick Start Ask the agent to analyze the blast radius of your current diff and prove the one fact it is safe because of by running real code.
