M87 Governed Bridge

Reference implementation of authority-separated multi-agent orchestration — the foundational research that became the M87 governance stack.

Status: Research artifact. For production governance tooling, see M87-Spine-lite and m87-audit-agent.

---

What It Demonstrates

An orchestration bridge connecting Claude and Gemini CLI agents with governance patterns applied at the handoff boundary — output validation, safe delivery, and structured task delegation between models.

The working proof that governance constraints can span model boundaries, not just single-agent sessions. The patterns here — authority-separated orchestration, output validation before handoff, fail-safe execution — became the M87 governance stack.

---

Quick Start

pip install -r requirements.txt
bash run_chain.sh prompt.txt

Customize CLAUDE.md and emotion_signal.json as needed.

---

Governance Patterns (Early Implementation)

• SPOT — Output policy violation detection
• FORT — Test integrity and output containment
• PARCEL — Safe output delivery before passing results between agents

---

Current Work

For production-ready governed agent infrastructure:

• M87-Spine-lite — Runtime governance for Claude Code. Policy-gated execution, cryptographic receipts, posture escalation.
• m87-audit-agent — Static code auditor. Deterministic enforcement, SHA-256 receipt chain, CI-ready.

---

License

MIT — © 2025 M87 Studio LLC.