spurpower-verification-before-completion

To stop AI from claiming code is complete without running tests, enforce a mandatory verification protocol that requires executing specific build or test commands and presenting output evidence before asserting success. This ensures all completion claims are backed by empirical validation.

To enforce empirical validation for bug fixes, mandate a verification gate where the agent must run the specific test case that previously failed, confirm it passes, and present the test output as proof before reporting the bug fix as complete.

A verification gate in code review is a strict sequence mandating the identification, execution, and review of test or build commands before any success claim. It ensures evidence-based reporting by requiring concrete output logs for all task completions.

No specific test frameworks are required. The verification protocol applies to any software development workflow by requiring the execution of your existing specific test or build commands and reading their output evidence prior to claiming work is done.

The best way to ensure evidence-based status updates is to apply a verification skill that mandates reading concrete output, such as test results or build logs, before allowing the agent to assert any task success. This guarantees reliability.