What problem does it solve? Security findings in Rust web services are easy to bury under architecture, error-handling, or test-review noise, and generic audits waste effort on out-of-scope concerns. This Skill performs a narrow, evidence-based review of seven specific security boundaries so concrete vulnerabilities surface with file/line proof instead of speculative warnings. ## Core Features & Use Cases - Boundary-scoped review: Covers secret exposure, authentication/authorization layering, untrusted-input sinks (SQL, paths, redirects, headers, shell), CORS/cookies, token comparison, rate limiting, and error-response leakage. - Evidence-first methodology: Requires tracing attacker-controlled sources to sensitive sinks before reporting, with severity assigned by impact and reachability, plus a separate Needs verification section for unresolved facts. - Optional scanner and references: Ships a ripgrep-based lead-generation script and boundary-specific reference checklists for secrets, input sinks, and browser/token/diagnostic concerns. - Use Case: Before merging a pull request that adds a new authenticated endpoint, run this review on the diff to confirm authorization is enforced at the service layer, tokens are compared in constant time, and error responses do not leak internals. ## Quick Start Ask the agent to review the current working changes with the rust-security-boundary-review skill and report findings with file locations without editing code.
