Tip: Working effectively with deep review

Recommendation

Run deep_review when you want security, quality, and test-gap findings in a single pass — not as three separate workflow calls.

DeepReviewAgentSDKWorkflow.execute() internally coordinates three specialized subagents (security-reviewer, quality-reviewer, and test-gap-reviewer) and synthesizes their output into one consolidated report. Calling each underlying workflow individually loses the cross-domain synthesis step.

Why

The orchestrator's final synthesis pass — not the subagent findings themselves — is where the prioritized, cross-domain Suggestions section is generated. Bypassing it means you never get the top 5–10 actionable next steps ordered by impact.

Tradeoff

A full deep-review run is slower than a single-domain review because all three subagents complete before synthesis begins. If you only need one dimension (for example, just security), prefer security-audit directly rather than filtering deep-review output after the fact.

Source files

Tags: review, security, quality, tests, comprehensive-review