verification-before-completion

To prevent premature task completion, enforce a strict verification protocol requiring the agent to run specific commands like test suites and analyze their output to confirm zero failures before declaring success.

An evidence before claims protocol is a verification process that mandates running testing or linting commands and reviewing their concrete output to substantiate any assertions of bug fixing or task completion.

You enforce mandatory verification by requiring the execution of specific testing or linting commands and analyzing their output to identify red flags or rationalizations before allowing any success declarations.

Yes, you can use this verification process for both linting and bug fixing workflows, as the protocol applies to all stages of work by requiring concrete command output as proof before making any claims.

Common red flags include rationalizations for skipping the verification process, such as assuming code changes are safe without running the test suite or presenting assertions without concrete command output as evidence.