flowbrain_core 0.1.2

FlowBrain Core — Judgment & Knowledge core for the MakeMind ecosystem. Two cooperating subsystems: Knowledge (5 facades over a 4-layer Fact/Skill/Profile/Philosophy structure plus Ops) and Agent (work [...]

example/README.md

FlowBrain Examples

Three runnable samples that exercise the package's public surface end-to-end against the built-in stub LLM — no external service or API key required.

File	Demonstrates
flowbrain_example.dart	Canonical hello — boot → create one worker agent → ask → read history → shutdown
multi_role.dart	The three AgentRole cooperating: manager routes a request to one of two worker candidates, then a reviewer evaluates the worker's reply
lifecycle_events.dart	Subscribing to system.eventBus.stream and observing the events emitted by create / ask / delete

Running

From the dart/ package root:

dart run example/flowbrain_example.dart
dart run example/multi_role.dart
dart run example/lifecycle_events.dart

Swapping in a real LLM

The examples use ModelSpec.stub(), which routes through StubLlmPort() — deterministic short replies, no network. To wire a real provider, supply llm: (or llmProviders: for multi-provider routing) when constructing the system:

final system = KnowledgeSystem.withAgents(
  llm: AnthropicLlmPort(apiKey: '...'),
);
final sara = await system.agents.createAgent(
  id: 'sara',
  displayName: 'Sara',
  role: AgentRole.worker,
  model: ModelSpec(provider: 'anthropic', model: 'claude-sonnet-4-6'),
  workspaceId: 'demo',
);

AnthropicLlmPort / OpenAiLlmPort / GeminiLlmPort etc. are provided by mcp_llm; any class implementing LlmPort works.