Architecture¶
Overview¶
The Agent Governance Toolkit provides deterministic application-layer interception โ every agent action is evaluated against policy before execution, at sub-millisecond latency. For high-security environments, composes with container/VM isolation for defense-in-depth.
Video Walkthrough Series¶
Community video series covering the toolkit architecture:
System Architecture¶
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
โ AGENT GOVERNANCE TOOLKIT โ
โ pip install agent-governance-toolkit[full] โ
โ โ
โ Agent Action โโโโบ POLICY CHECK โโโโบ Allow / Deny (< 0.1 ms) โ
โ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ AGENT OS ENGINE โโโโโโบโ AGENTMESH โ โ
โ โ โ โ โ โ
โ โ โ Policy Engine โ โ โ Zero-Trust Identity โ โ
โ โ โ Capability Model โ โ โ Ed25519 / SPIFFE Certs โ โ
โ โ โ Audit Logging โ โ โ Trust Scoring (0-1000) โ โ
โ โ โ Action Interception โ โ โ A2A + MCP Protocol Bridge โ โ
โ โโโโโโโโโโโโโโฌโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโโโฌโโโโโโโโโโโโโโโ โ
โ โ โ โ
โ โผ โผ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ AGENT RUNTIME โ โ AGENT SRE โ โ
โ โ โ โ โ โ
โ โ โ Execution Rings โ โ โ SLO Engine + Error Budgetsโ โ
โ โ โ Resource Limits โ โ โ Replay & Chaos Testing โ โ
โ โ โ Runtime Sandboxing โ โ โ Progressive Delivery โ โ
โ โ โ Termination Control โ โ โ Circuit Breakers โ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ AGENT MARKETPLACE โ โ AGENT LIGHTNING โ โ
โ โ โ โ โ โ
โ โ โ Plugin Discovery โ โ โ RL Training Governance โ โ
โ โ โ Signing & Verificationโ โ โ Policy Rewards โ โ
โ โโโโโโโโโโโโโโโโโโโโโโโโโโโโ โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ โ
โ โ
โโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโโ
Security Model & Boundaries¶
| Enforcement Capability | Defense-in-Depth Composition |
|---|---|
| Intercepts and evaluates every agent action before execution | Add container isolation (Docker, gVisor, Kata) for OS-level separation |
| Enforces capability-based least-privilege policies | Add network policies for cross-agent communication control |
| Provides cryptographic agent identity (Ed25519) | Add external PKI for certificate lifecycle management |
| Maintains append-only audit logs with hash chains | Add external append-only sink (Azure Monitor, write-once storage) for tamper-evidence |
| Terminates non-compliant agents via signal system | Add OS-level process.kill() for isolated agent processes |
The POSIX metaphor (kernel, signals, syscalls) is an architectural pattern โ it provides a familiar, well-understood mental model for agent governance. The enforcement boundary is the Python interpreter, which is the same trust boundary used by every Python-based agent framework (LangChain, AutoGen, CrewAI, OpenAI Agents SDK).
Production recommendation: For high-security deployments, run each agent in a separate container with the governance middleware inside. This gives you both application-level policy enforcement and OS-level isolation.
Trust Score Algorithm¶
AgentMesh assigns trust scores on a 0โ1000 scale with the following tiers:
| Score Range | Tier | Meaning |
|---|---|---|
| 900โ1000 | Verified Partner | Cryptographically verified, long-term trusted |
| 700โ899 | Trusted | Established track record, elevated privileges |
| 500โ699 | Standard | Default for new agents with valid identity |
| 300โ499 | Probationary | Limited privileges, under observation |
| 0โ299 | Untrusted | Restricted to read-only or blocked |
Default score for new agents: 500 (Standard tier). Score changes are driven by policy compliance history, successful task completions, and trust boundary violations. Full algorithm documentation is in agent-governance-python/agent-mesh/docs/TRUST-SCORING.md.
Benchmark Methodology¶
Policy enforcement benchmarks are measured on a 30-scenario test suite covering the OWASP Agentic Top 10 risk categories. Results (e.g., policy violation rates, latency) are specific to this test suite and should not be interpreted as universal guarantees. See agent-governance-python/agent-os/modules/control-plane/benchmark/ for methodology, datasets, and reproduction instructions.
Full benchmark results: BENCHMARKS.md