09 — Topology A (EOAM) — Technical Minimal Demo

Goal: One market tick drives a single OBSERVATION on an in-memory EventBus (create_event_bus). Two reactive agent classes defined in run.py subscribe by instrument scope, each publishes a POLICY_PROPOSAL, the runtime arbitrates with a fixed priority_matrix, then DIM validates the winner through DecisionRuntime.evaluate_proposal after handshake — all with memory_storage() and no samples/shared, no YAML, no LLM. QuoteGenerator lives under mocks/ (same tick shape as samples/31_finance_trading). This layout follows .cursor/rules/06-technical-sample-development-guide.mdc.

Mechanisms: DecisionRuntime, memory_storage, create_event_bus, EventType.OBSERVATION, scope dispatch, priority arbitration, IntentRetryGovernor, DIM with default in-memory proposal audit rows.

---

Architecture / flow

---
title: EOAM technical minimal — one observation cycle
config:
  layout: elk
  theme: neutral
  look: classic
---
flowchart LR
  classDef userSpace fill:#E8EAF6,stroke:#3F51B5,stroke-width:2px,color:#1A237E,font-weight:bold;
  classDef kernelSpace fill:#E8F5E9,stroke:#388E3C,stroke-width:2px,color:#1B5E20,font-weight:bold;
  classDef wall fill:#FFF3E0,stroke:#F57C00,stroke-width:2px,color:#E65100,font-weight:bold;

  subgraph US[User space in run.py]
    QG[QuoteGenerator]:::userSpace
    RA[RiskAgent]:::userSpace
    SA[StrategyAgent]:::userSpace
    BUS[EventBus]:::userSpace
  end

  subgraph W[The Wall]
    DIM[evaluate_proposal]:::wall
  end

  subgraph KS[Kernel space]
    RT[DecisionRuntime]:::kernelSpace
  end

  QG -->|OBSERVATION| BUS
  BUS --> RA
  BUS --> SA
  RA -->|POLICY_PROPOSAL| BUS
  SA -->|POLICY_PROPOSAL| BUS
  BUS -->|winner| DIM
  DIM --> RT

---

How to run

From the repository root:

python samples/09_topology_a_eoam/run.py

No API keys or database files are required. Tweak constants at the top of run.py (_PRIORITY_MATRIX, _INSTRUMENT, _QUOTE_SEED, wake-up flags) to experiment.

---

Expected output

You should see two handshake log lines, one OBSERVATION dispatch, two POLICY_PROPOSAL dispatches (when with_logging is false on the bus, only kernel logs appear), DIM verdict, mock execution on ACCEPT, a SUMMARY log line, and the final [SUMMARY] print:

INFO Handshake: agent_id=agent_risk ver=1.0.0 accepted
INFO Handshake: agent_id=agent_strategy ver=1.0.0 accepted
INFO [DFID=...] EOAM: publishing OBSERVATION for BTC-USD
INFO [DFID=...] DIM verdict=ACCEPT reason=VALIDATION_PASSED
INFO [DFID=...] Mock execution for agent_risk policy=ALERT
INFO [DFID=...] SUMMARY proposals=2 chosen_agent=agent_risk chosen_policy=ALERT verdict=ACCEPT executed=True

[SUMMARY] DFID=... proposals=2 chosen=agent_risk policy=ALERT verdict=ACCEPT executed=True

With default thresholds, Risk usually wins ALERT because quote volatility exceeds _RISK_VOL_ALERT_ABOVE.