verification-before-completion

To prevent AI from claiming code is fixed without running tests, enforce a mandatory verification gate that requires executing full test suites and reviewing exit codes before allowing any positive status assertions. This blocks false completion claims by demanding fresh evidence.

Commit validation is the process of running full verification commands and confirming zero test failures before creating a pull request. It is needed to prevent shipping broken code to production by ensuring claims of readiness are backed by observed evidence.

To stop an AI agent from rationalizing skipping test execution, apply counterarguments against common excuses like 'I'm confident' or 'just this once'. Enforcing a strict 5-step verification process ensures the agent runs commands and reviews complete output before claiming completion.

Yes, you can use commit validation for build verification and bug fix validation. The process applies to software development workflows by requiring fresh verification evidence, reviewing complete output, and aligning claims strictly with observed test execution results before any positive assertions.

Common invalid completion claims during code commits include stating 'tests should pass' without running them or asserting a bug is fixed without evidence. Mandatory verification prevents these by listing frequent invalid claims and enforcing fresh execution output before any completion status is declared.