$ python main.py --mock

[INFO] Starting simulation campaign (mock=True, model=gpt-4o)

# 1. CONTROL PLANE
[INFO] LLM call by campaign_planner | spent: $0.00 / $10.00
[INFO] LLM call by campaign_planner complete | cost: $0.03
[INFO] LLM call by persona_generator x3 clusters | cost: $0.09
[INFO] Generated 6 personas across 3 clusters
[INFO] Static analysis flagged 7 endpoints
[INFO] Dynamic analysis flagged 7 endpoints
[INFO] LLM call by vuln_scan_agent | cost: $0.03

# 2. EXECUTION PLANE
[INFO] LLM call by journey_generator x6 journeys | cost: $0.18
[INFO] Generated 6 journeys
[INFO] Journey j1: CHAOS_TRIGGERED (6 steps, 24ms)
[INFO] Journey j2: CHAOS_TRIGGERED (5 steps, 33ms)
...
[INFO] Collected 6 simulation results

# 3. OBSERVATION PLANE
[INFO] Kafka stream processor initialized
[INFO] ObserverAgent identified 12 rule-based anomalies
[INFO] AnomalyDetectorAgent enriched 12 anomalies (3 statistical, 2 LLM semantic)
[INFO] EdgeCaseAgent discovered 4 novel edge cases

# 4. REPORTING PLANE
[INFO] ResilienceScorerAgent calculated score: 85/100
[INFO] ReportGeneratorAgent created campaign report
[INFO] QABridgeAgent pushed 4 edge cases to QA pipeline

[INFO] Campaign completed in 120 seconds.

.github/                       # GitHub Actions workflows
├── workflows/
│   └── ci.yml
├── docs/                        # Documentation and diagrams
├── agents/                      # Core agent implementations
│   ├── control/
│   │   ├── campaign_planner.py
│   │   ├── persona_generator.py
│   │   └── vuln_scan_agent.py
│   ├── execution/
│   │   ├── journey_generator.py
│   │   ├── simulator.py
│   │   └── chaos_agent.py
│   ├── observation/
│   │   ├── observer.py
│   │   ├── anomaly_detector.py
│   │   └── edge_case_agent.py
│   └── reporting/
│       ├── resilience_scorer.py
│       ├── report_generator.py
│       └── qa_bridge_agent.py
├── persona/
│   ├── clustering.py               # K-Means + elbow method + PII scrubber
│   ├── models.py                   # PersonaPopulation + diversity scoring
│   └── mutator.py                  # Unicode/boundary/timing mutations
├── chaos/
│   ├── vulnerability_map.py        # 3-stage OpenAPI + metrics analysis
│   ├── injection_engine.py         # mitmproxy + MockChaosEngine fallback
│   └── budget_governor.py          # Redis atomic counters + auto-halt
├── observation/
│   ├── stream_processor.py         # Kafka consumer + mock list fallback
│   ├── anomaly_rules.py            # Layer 1 rule definitions
│   └── edge_case_discovery.py      # DBSCAN + novelty scoring
├── persistence/
│   ├── campaign_db.py              # PostgreSQL/JSONB -> SQLite fallback
│   └── knowledge_base.py           # pgvector -> numpy fallback
├── llm/
│   ├── prompts.py                  # 12 prompt templates (6 system/user pairs)
│   └── factory.py                  # LLM factory + CostGovernor wrapper
├── observability/
│   ├── metrics.py                  # Prometheus counters/histograms/gauges
│   └── tracing.py                  # OpenTelemetry setup
├── api/app.py                      # FastAPI (POST /api/campaigns, GET /health)
├── main.py                         # CLI: --mock, --server, --output
├── requirements.txt
└── tests/
    ├── unit/                       # 24 unit tests
    └── integration/                # 3 integration tests (full pipeline)