verification-before-completion

To run validation checks before completion, inspect project context files like README and package.json to create a validation plan, execute deterministic tests, type checks, and builds, then report outcomes with supporting evidence.

Verifying code changes against acceptance criteria requires creating a validation plan that prioritizes high-signal tests for changed areas, running deterministic checks like linting and builds, and distinguishing passing checks from unresolved risks.

A validation plan identifies required test commands by inspecting project context files such as Makefile, pyproject.toml, package.json, and CI configurations to determine the deterministic checks needed for completion.

You can use this validation process for documentation tasks and software development alike, as it applies to any task work featuring explicit acceptance criteria that require deterministic checks and manual verifications before completion.

A completion report with failing checks or remaining risks must clarify unresolved assumptions and distinguish passing checks from failures, producing evidence of the validation steps executed rather than simply marking the work done.