MycoMiner: Multi-Agent Intelligence System
Human-crafted mushroom intelligence with governance and explainability
Agent Console
Checking...Welcome to MycoMiner Agent Console
Start a conversation with the multi-agent system
Governance Log
0 total • 0 completed • 0 failed • 0 pending
No logs available
Showing logs from all sessions
Multi-Agent System Features
Multi-Agent Orchestration
Intelligent routing to specialized agents (Reasoner, Planner, Critic, General) based on message intent with deterministic scoring.
Policy Governance
Configurable pre and post-message validation with structured error reporting and severity levels (info/warning/error/critical).
Governance Logging
Complete audit trail with filterable logs, timestamps, agent metadata, and performance metrics. Exportable for compliance.
Semantic Context
Rolling conversation windows with session isolation. Maintains up to 50 messages per session across 100 concurrent sessions.
Explainability Graph
Visual reasoning paths with typed nodes (policy/routing/agent/response) and metadata-enriched edges. Auto-prunes at 300 nodes.
Real-Time Monitoring
Live updates with manual refresh controls, auto-refresh toggles, error handling, and loading states for all panels.
System Architecture
User Message: Enters through AgentConsole, creates governance log entry
Pre-Policy Check: Validates message length, content safety, rate limits
Semantic Context: Adds message to rolling context window, builds context string
Orchestration: Routes to specialized agent using keyword/pattern matching with scoring
Agent Execution: Selected agent processes with enriched context and system prompt
Post-Policy Check: Validates response quality, length, content safety
Graph & Logs: Updates explainability graph nodes/edges, completes governance log
Response: Returns to user with full metadata (routing, policies, graph stats)