What problem does it solve? Code reviews often miss breakage that grep and symbol search cannot find, such as changes to API responses, wire formats, feature flags, or downstream consumers. This Skill identifies the real risks of a change before it ships and proves the key safety assumption by executing actual code instead of trusting a convincing writeup. ## Core Features & Use Cases - Hidden Breakage Detection: Traces risks that symbol search misses, including JSON payloads, DB columns, wire formats, cross-language consumers, and code several hops downstream. - Proof Over Prose: Ranks each safety fact on a five-level confidence ladder and requires running a script or test against the real code before calling anything proven. - Honest Risk Reporting: Separates confirmed risks with file:line citations from checked-and-cleared cases, and marks anything unproven instead of rounding up. - Use Case: Before merging a diff that touches a shared cache invalidation function, use this Skill to find the one fact the change is safe because of, write a small script that calls the real function, and ship the proof alongside the risk list. ## Quick Start Ask the assistant to analyze the blast radius of your current diff and prove the one fact it is safe because of by running real code.
