coding-worker-agent

To execute a bounded coding task with evidence-first completion, define a task with explicit scope and request completion that includes touched files, artifacts, and validation results. This enforces strict scope discipline and produces traceable artifacts before acceptance.

Evidence-first completion for coding and refactor tasks is a verification mechanism that requires truthful progress reporting. It highlights missing validation and open risks before handoff, ensuring all touched files and artifacts are traceable and validated.

To prevent scope creep during coding, fixes, and refactoring, apply strict scope discipline by executing one bounded task at a time. This ensures the implementation stays within explicit boundaries and satisfies requirements for clear scope.

Yes, you can use this approach for wiring and implementation tasks requiring validation. It applies to any coding task where the scope is explicit and verification is required before acceptance, producing validation evidence and highlighting open risks.

Bounded coding tasks for complex refactors are limited to one task at a time with explicit scope. If validation is missing, the mechanism highlights open risks before handoff rather than auto-resolving them, requiring manual intervention for acceptance.

Your coding task requires traceable artifacts before acceptance to ensure truthful progress reporting. This evidence-first approach validates touched files and changes, preventing unverified implementations from passing without explicit validation results.