openspec-sp-verify-before-complete

Test verification before task completion is enforced by requiring concrete evidence such as passing test outputs or successful builds. You must run validation commands, read the outputs, and present verifiable proof rather than making vague or performative completion statements.

Evidence-first completion for bug fixes and feature merges mandates that assertions of completion are backed by readable verification outputs. This prevents premature claims by requiring successful builds, reproduced fixes, or regression test results before reporting work as done.

To stop automated agents from making unsubstantiated completion claims, apply communication guardrails that block vague language and enforce precise reporting formats. Require agents to run validation commands and confirm evidence for assertions before reporting task completion.

Yes, this verification approach supports multiple validation methods including build and compile success checks, bug reproduction, regression testing, and independent automated checks. It satisfies requirements to run validation commands and read outputs for software development scenarios.

The best way to verify code changes before closing a bug ticket is to run the chosen verification method, read and record the outputs, and report concrete evidence. This ensures the recipient receives verifiable proof of the regression test or fix reproduction.

You need to require proof of passing tests for software development handoffs when you want to ensure the recipient receives verifiable proof rather than assertions. This prevents premature or unsubstantiated completion claims during work transitions.