What problem does it solve? Code reviews often miss breakage that lives outside the diff: downstream callers, wire formats, DB columns, feature flags, and library behavior that grep cannot reveal. This Skill finds what a change could break somewhere else before it ships, and instead of trusting a convincing writeup, it proves the one safety fact the change depends on by running real code. ## Core Features & Use Cases - Hidden Breakage Discovery: Traces risks grep misses, including JSON API responses, DB columns, wire formats, cross-language consumers, feature flags, and code three hops downstream. - Evidence-Ranked Safety Facts: Ranks each safety claim on a five-level confidence ladder, from assertion up to reproduction in the running app, and marks anything unproven honestly. - Proof by Execution: Writes and runs a small script or test that calls the real code and fails loudly if the safety assumption is wrong. - Use Case: Before merging a diff that changes a cache-invalidation call, ask for its blast radius to identify the single fact that makes it safe, verify the pinned library version's actual behavior, and get a runnable script that proves it. ## Quick Start Ask the agent to analyze the blast radius of your current diff or a specific change and prove the key safety assumption with a runnable test.
