kramme:verification-before-completion

Enforce honest completion status by requiring fresh verification evidence before any success claim. This means defining a verification command, executing it fully, evaluating the output, and only then reporting the result as auditable evidence.

The best way to ensure PRs and builds are backed by verifiable output is applying a repeatable verification workflow that identifies the command, runs it, reads the output, verifies the result, and reports it before declaring work complete.

Enforce a repeatable verification workflow by identifying a specific verification command, running the full command execution, reading and evaluating the output, and reporting the verified result before allowing any task closure or completion claim.

Yes, this verification process works for task closures, tests, PRs, and builds without external dependencies. It operates by requiring a defined verification command, full execution, output evaluation, and result reporting to ensure claims are supported.

You need to run a verification command before claiming task completion because it enforces honesty in completion status by requiring fresh verification evidence. Claims must be supported by actual verifiable command output rather than assumptions.

The limitation is that this verification approach requires a clearly defined verification command for every task. Without a runnable command to execute and evaluate, the process cannot generate the fresh evidence needed to support a completion claim.