Skip to content

Testing And Validation

Purpose: define the evidence needed before a team trusts agent-assisted work.

Use this section by itself when the team needs stronger acceptance criteria, tests, review gates, or release evidence.

Inputs

  • Behavior being changed.
  • Test suite and known gaps.
  • Manual review path.
  • Environments.
  • Acceptance criteria.
  • Failure and rollback path.

Outputs

  • Test strategy.
  • Review checklist.
  • Evidence to capture.
  • Failure interpretation.
  • Golden scenario suite.
  • UAT scripts.
  • Readiness gaps.

Agent Task Contract

Create a validation plan that connects each workflow or change type to tests, manual checks, acceptance evidence, and recovery.

Help update test plans, acceptance criteria, review gates, evidence checklists, and failure interpretation as the work changes. Keep human review and acceptance explicit.

Do not treat generated tests or agent review as enough by themselves.