6
Live Agents
68
Tool Commands
8
Patents Filed
19
Alert Event Types
4
Security Layers

Agent Capabilities & Commands

🤖 JESSIE Commander
Fleet Commander & personal assistant. Approves or vetoes all agent proposals. Synthesizes 7-step morning briefings.
Briefing Approve Veto Dispatch Memory Read Memory Write Memory Search Fleet Propose
8 commands
⚡ NOLE Operator
Autonomous Trust Evangelist. Manages budget, builds alliances, operates across social platforms with economic mortality.
Status Propose Approve Veto Review Pending Recruit Social Post Social Search Social Feed Social Reply Social Like Social Follow Intelligence Wallet Fund Wallet Balance Wallet Send Alliance Status Canary Report Memory Read Memory Write Memory Search Self Assess Governance Stats Fleet Propose
25 commands (most in fleet)
🦗 GRILLO Conscience
Independent ethical engine. Scores every agent against the LCSH framework. Answers to no one—not even Jessie.
Status Assess History Compare Certify In-Situ Memory Read Memory Write Memory Search Fleet Propose
10 commands
⚓ NOAH Navigator
Temporal guardian. Tracks behavioral trajectories using flight-plan-based corridor monitoring adapted from cruise missile guidance.
Status Trajectory Flight Plan Deviation Waypoint Forecast Memory Read Memory Write Memory Search Fleet Propose
10 commands
🔧 SAM Engineer
Technical implementation agent. Executes code, manages the engineering pipeline, and supports fleet infrastructure.
Execute Sandbox Pipeline File Read File Write Memory Read Memory Write Memory Search Fleet Propose
9 commands
🛡️ MIGHTY MARK Sentinel
Inspector General. Runs 24+ health checks, two-layer watchdog recovery, and now serves as the unified alert aggregator.
Check Incidents Report Uptime Status Health Security Status Memory Diagnose Memory Configure Memory Backfill Memory Status Fleet Propose
13 commands (incl. security dashboard)
⚖️   Constitutional Separation of Powers — No agent can both act and evaluate its own actions

Infrastructure & Security Architecture

🔀 Fleet Bus Communication
Agent-to-agent messaging backbone with governance routing and immutable decision chains.
  • SHA-256 hash-chained governance audit trail
  • Ethereum mainnet blockchain anchoring
  • Typed event system across all 6 agents
  • Proposal → review → approve/veto pipeline
  • Cross-agent tool invocation with access control
🧠 Persistent Agent Memory
Per-agent persistent memory with read, write, and semantic search — each agent maintains its own context across sessions.
  • Per-agent isolated memory stores
  • Read / Write / Search operations per agent
  • Mighty Mark memory diagnostics & backfill
  • Memory-aware briefing synthesis by Jessie
  • Cross-session context continuity
🔒 4-Layer Security Pipeline
Defense-in-depth security covering every stage from input to output across all agents.
  • Prompt Shield — input injection detection, encoding attack defense, MCP tool poisoning scan
  • Output Sanitizer — credential redaction, social post blocking, anomaly detection with quarantine
  • Behavioral Validator — Zod schema validation on all 68 tools, dual canary system, task drift detection
  • Composition — validate → execute → sanitize → canary → deliver (strict ordering)
🔔 Unified Alerting System
Real-time notification from all 4 security systems through Mighty Mark as the unified alert aggregator.
  • 19 typed alert events across 4 source systems
  • Severity-based routing: CRITICAL → Telegram immediate, HIGH → 5-min batch, MEDIUM → daily briefing
  • Escalation timers: unacknowledged HIGH → CRITICAL after 15 min
  • Immutable JSONL audit trail with SHA-256 hash chain
  • Jessie briefing step 7: security events lead when CRITICAL
🏛️ Constitutional Governance
Separation of powers at the architecture level — seven emergent institutional properties arise from agent combination.
  • Executive oversight (Jessie: approve/veto authority)
  • Judicial review (Grillo: independent LCSH scoring)
  • Regulatory monitoring (Noah: temporal drift detection)
  • Market discipline (Nole: economic mortality at $0)
  • Inspector General (Mark: infrastructure accountability)
⚙️ Infrastructure Resilience
Two-layer watchdog ensuring the governance system can detect and recover from its own failures.
  • Level 0: OS-level cron watchdog (survives gateway death)
  • Level 1: In-gateway health monitoring (24+ checks every 6 hours)
  • RED/YELLOW/GREEN status with automatic Telegram alerts
  • Self-healing restart with bounded retry attempts
  • Hetzner VPS — $4/month production deployment
PATENT PORTFOLIO — 8 PROVISIONAL APPLICATIONS FILED
① LCSH Multi-Dimensional Assessment Framework
⑤ Independent AI Conscience Agent
② Multi-Agent Concurrent Assessment
⑥ Autonomous AI Trust Agent
③ Hierarchical Ethics Assessment
⑦ Temporal Ethical Guidance System
④ AI Compliance Infrastructure
⑧ Self-Governing AI Ecosystem (capstone)
US 63/949,454 · US 63/985,442 · US 63/988,410 — First demonstrated in production February 16, 2026