setup-evaluation

To validate process decompositions and architecture designs, an independent evaluator checks step coverage, handoff protocols, and topology alignment. This setup evaluation reads your markdown artifacts to detect orphan steps and structural mismatches, producing a PASS or FAIL report with remediation guidance.

Multi-agent workflows fail from undefined handoffs and orphan steps when process and architecture specs lack independent review. A setup evaluation catches these structural errors and topology mismatches before execution, preventing wasted agent runs and costly deployment failures.

The best way to review agent-chain orchestration for topology mismatches is running an independent setup-evaluator agent. It verifies topology alignment, agent boundaries, and handoff protocols against your architecture specifications, avoiding the confirmation bias from builders evaluating their own designs.

Yes, you can evaluate project setup markdown artifacts for parallelism and tool availability. The setup evaluation verifies step coverage and tool availability within your process decomposition, ensuring all steps are assigned and execution requirements are met before returning a verdict.

Yes, architecture review for agent workflows requires separate process and architecture markdown artifacts. The evaluator reads both to check step coverage, handoff protocols, and topology alignment, ensuring process entries and architecture specifications match before execution.

When repeated architecture reviews fail to resolve structural errors, the setup evaluation escalates to the user. After producing multiple FAIL reports with enumerated issues and remediation guidance, the evaluator forwards the unresolved handoff protocols and topology mismatches for manual intervention.