HyperoruDocumentation
Workflows

Multi-agent system

The eleven specialist roles, typed artifact handoffs, activation rules, guardrails, model tiers, and adjudication flow.

Agents collaborate through validated artifacts—not hidden chain-of-thought. Every artifact records its agent definition version, model, context hash, evidence IDs, confidence, usage, latency, and decision summary.

01
CoordinateAudit Coordinator

scope + stage decisions

02
ContextContext Curator

bounded evidence packs

03
VerifyEvidence Verifier

citation + contradiction checks

04
AnalyzeSecurity Engineer

technical findings

05
ModelSecurity Architect

boundaries + structural risk

06
SynthesizeReport Adjudicator

canonical result

Role registry

RolePrimary responsibilityTypical tier
Audit Coordinatorscope, activation, stage and retry decisionsLuna
Context Curatorassemble bounded ContextPacksLuna
Evidence Verifierreject unsupported claims and contradictionsLuna
Security Engineertechnical vulnerabilities and control failuresTerra
Security Architectboundaries, reachability, blast radius, structural weaknessTerra
Cloud/IaC Analystinfrastructure entities, policy, network, identityTerra
CI/CD Analystpipeline, artifact, deployment, and provenance relationshipsTerra
Threat Modelerassets, actors, paths, abuse cases, mitigationsTerra
Remediation Engineerdeterministic patch proposal and static rescan planTerra
Diagram Architectsafe views of the canonical graphTerra
Report Synthesizer / Adjudicatorreconcile artifacts into canonical JSONTerra

Sol-class models are reserved for an explicitly approved deep review. They are never selected merely because another agent failed.

YAML definitions

Definitions under backend/llm/prompts, backend/llm/skills, and backend/llm/workflows are versioned and validated at startup. Each definition declares:

  • role and activation rules;
  • model tier and bounded tools;
  • Pydantic input and output schemas;
  • token, latency, and cost limits;
  • memory policy and evidence requirements;
  • prompt-injection, secret, and unsupported-claim guardrails.

Unknown tools, invalid schemas, unsafe dynamic imports, or missing versions fail startup and CI.

Evidence adjudication

An observation becomes canonical only after the verifier can resolve every cited evidence ID in the same audit and workspace. Unsupported claims remain hypotheses. Conflicting agent artifacts are retained and sent to adjudication with the underlying evidence—not resolved by majority vote.

On this page