aep-gen-eval

A generator/evaluator pattern for artifact validation separates production from quality assessment by assigning distinct agent roles to generate work and evaluate it. This ensures honest quality checks by using structured intermediates for communication between the evaluator and generator.

Automating design validation and document assessments is best handled by running a gen/eval loop using a dedicated scoring framework and evaluation protocol. This process evaluates artifacts directly against structured criteria to ensure honest quality checks.

Yes, you can use agent contracts to enforce skill assessment protocols by defining specific generator and evaluator roles. This requires separate agents to handle generation and evaluation, communicating via structured intermediates to validate artifacts.

Artifact validation using a generator/evaluator pattern requires separate agent roles for the generator and evaluator. This separation is necessary to ensure honest quality checks during code reviews, design validations, and document assessments.

You can integrate a scoring framework with other build and validate skills like /aep-build and /aep-validate. The evaluation protocol easily integrates into existing workflows to run a gen/eval loop on any new code change or artifact.