What happened
Drata announced (Aug 4, 2026) Limited Availability of AI Agent Governance, built on a Drata Sensor, MCP Proxy, and tamper-evident telemetry, to discover, monitor, and govern AI agents inline (stopping violating actions before execution); ships deepest for Anthropic today, with OpenAI/Vertex AI/Bedrock coverage in development.
Why it matters
Extends a compliance-automation platform into real-time agent governance tied directly to EU AI Act, AIUC-1 and ISO 42001 evidence requirements, explicitly citing the Anthropic and Hugging Face agent incidents as the threat model.
Applicability
Compliance and security teams using Drata for GRC should evaluate for Anthropic-based agent fleets now; broader coverage still pending.