verification-loop

Fresh build and test verification is enforced by executing build, test, and lint commands directly in the active session and recording outputs with timestamps. This prevents developers from using cached results or past runs to claim task completion.

Evidence-based task completion is a workflow mechanism requiring validation command execution in the current session alongside captured outputs and timestamps. It standardizes verification to prove builds, tests, and lint passes without regressions.

To verify builds without cached results, execute environment-specific commands like npm run build && npm test && npm run lint for JavaScript, or activate venv and run pytest && flake8 for Python, ensuring the active session captures fresh output evidence.

Yes, the verification loop works with npm and pytest environments by executing relevant commands in the active session. It supports JavaScript and TypeScript projects using npm, and Python projects using pytest and flake8 for validation.

Recording timestamps when running lint and test validations supports auditability by proving the results are fresh. This prevents premature task completion claims based on past runs, cached results, or outdated assumptions from previous sessions.

Session-based verification is limited to the active session and cannot use past runs or cached results to claim completion. It requires executing commands directly in the current environment, meaning it does not validate work performed outside the tracked session.