synthesizing-review-findings
synthesizing-review-findings
When to use
Merges the findings of several code-quality-atlas lenses — and of any other review method run alongside them (the built-in code-review skill, a framework review like BMAD, linter output, or human notes) — into one review: deduplicates issues raised by more than one source, reconciles lenses that pull opposite ways (e.g. restraint vs. coverage, cache vs. profile), ranks by severity, and ends on a single block/approve verdict. Use after running any set of atlas review lenses (and any companion reviewers) on a change — typically 3-8 in standard review mode, uncapped in comprehensive mode — when assembling multi-source review output into one report, or when overlapping findings need deduplicating and prioritizing.
Shape: composition. Runs after choosing-review-lenses has picked the lenses and you have each lens's findings in hand; it produces the single review a human or agent actually reads. It adds no new checks of its own — it only merges.
Fan-out model
Fan-out is advisory by default: you run each lens the router named, collect its findings, then apply the steps below to merge them. The finding shape is fixed (see Finding contract) so a harness that can invoke lenses in parallel may mechanize the same merge — the dedupe and ranking rules are deterministic. Automated or by hand, the output is identical. The same fixed finding shape also lets an orchestrator fan out across many repositories — one agent per repo emitting findings in this contract — and aggregate them centrally (see the multi-repo runbook under Going deeper).