verification-before-completion

Enforce verification before completion by mandating explicit command execution and output checks. This prevents premature success assertions by requiring test suites to pass and build commands to exit successfully before reporting status.

The best way to ensure build verification is applying an evidence before claims policy. Run specific commands, confirm zero test failures, and verify the build exits with code zero before declaring any task complete.

Stop false status reporting by enforcing strict verification checks during bug fixing. Require the agent to execute commands and validate outputs, preventing rationalizations that lead to shipping incomplete or broken work.

Yes, you can enforce linting and test execution before completion. The verification policy applies to all development stages, requiring commands to run and outputs to be checked before any task success is claimed.

Agents claim completion without running builds due to common rationalizations and red flag behaviors. Enforcing an evidence before claims policy mandates explicit command execution to prevent these premature assertions of success.