Enforces a strict evidence-based protocol to verify tasks, bug fixes, and builds before claiming completion.
The Verification Guard skill establishes a mandatory 'verification gate' to ensure high-integrity software development. It prevents Claude from making unverified claims of success by requiring a structured checklist: identifying the correct verification command, executing it freshly, reading the full output, and confirming the result with evidence. By eliminating hedging language like 'should work' and 'probably fixed,' this skill builds developer trust and ensures that every code change is backed by passing tests, clean builds, and validated logic before any work is considered done.
Key Features
01Independent verification of automated tool and agent reports
02Red-green regression test validation to prove bug fixes
0362 GitHub stars
04Elimination of hedging language and unverified assertions
05Identification of specific commands needed to prove success
06Mandatory verification checklist for every completion claim
Use Cases
01Confirming a bug fix with a regression test that fails before the fix and passes after
02Validating that complex agent-led refactors have actually updated the codebase correctly
03Ensuring a build or linter is 100% clean before committing changes or opening a PR
