Insights on AI agent governance
Thinking about accountability, compliance, and the governance infrastructure AI agents need.
Orchestrating Accountability: Governance Strategies for Multi-Agent Systems and Complex Autonomous Loops
In a Multi-Agent System, the greatest risk isn't a single agent going rogue — it's the unintended chemistry of their interactions. Part 4 of the agentic governance series moves beyond individual guardrails to the global orchestration layer, solving Intent Integrity and Trust Propagation in autonomous loops.
The Identity Gap: Why OAuth Was Never Built for Agents
OAuth was designed for humans clicking buttons in browsers. Autonomous agents need short-lived cryptographic identities, scoped delegation, and policy-bound context — not master keys. Here's the architecture that closes the gap.
A Deep Dive into Policy-as-Code: Building Real-Time Permission Slips for Your AI Agents
How to use OPA and Cedar to give your agents machine-readable, hot-swappable permissions that enforce in microseconds — and produce the audit trail regulators actually want.
Logging Isn't Auditing: Moving from Agent Observability to True Accountability
Agent observability answers what happened. Accountability answers why it was allowed and who authorized it. Here's why that distinction matters — and what to do about it.