Multi-agent architecture: security, performance, API, and style agents run in parallel. An aggregator agent merges outputs using strict schemas. Full trace for audit and debugging.
{
"merged_findings": [
{
"file": "string",
"line_range": [number, number],
"severity": "critical" | "high" | "medium" | "low",
"category": "security" | "performance" | "api" | "style",
"source_agent": "string",
"title": "string",
"comment": "string",
"confidence": number,
"trace_id": "uuid"
}
],
"conflicts_resolved": [
{
"agents": ["string", "string"],
"resolution": "string",
"rationale": "string"
}
],
"aggregate_risk_score": number
}Reserve for high-value situations where comprehensive coverage justifies the cost