validator

To verify code changes actually satisfy task requirements, you can validate task files against actual diffs and test results. This process maps concrete checklist items to the specific implementation code to catch disconnected or partial work.

Yes, you can automatically run relevant tests and append a validation report to a task file. The validation process executes the most relevant tests, records whether the implementation holds up, and appends a structured summary directly into the task file.

Task validation does work with backend, frontend, API, prompt, and integration review workflows. It applies to repository tasks added or updated in task_*.md, tasks/task-*.md, and docs/tasks/*.md file formats to inspect connected files and implementation paths.

The best way to detect regressions after updating a feature branch is to review diffs, map requirements to the implementation, and run relevant tests. This confirms the task is actually done and identifies any gaps in the updated code.

To map task checklist items to actual code changes, you break the task into concrete requirements and inspect the connected files and implementation paths. This change inspection process catches partial work by mapping each item to the satisfying code.