AgentsAugust 16, 2026via InfoQ AI/ML
AWS Open-Sources Dogwood, Extending Cedar to Govern Sequences of Agent Tool Calls
Why it matters
Agent governance is moving from per-request rules to stateful, multi-step reasoning. Dogwood extends Cedar to track agent behavior across sequences of actions—approvals, rate limits, running totals—addressing a critical reliability and control gap as agents move into production workflows.
Key signals
- AWS open-sources Dogwood under Apache 2.0
- Extends Cedar policy language with temporal conditions
- Governs sequences of agent tool calls, not single requests
- Covers approvals, rate limits, running totals
- Integrated into AgentCore Policy
- Reference interpreter not production-ready
- Addresses stateful agent governance for multi-step workflows
The hook
AWS open-sources Dogwood: agent policy language that remembers what agents did yesterday. Temporal governance for sequences of tool calls, not just single requests.
AWS has open-sourced Dogwood, a policy language extending Cedar with temporal conditions so rules can reason about an agent's prior tool calls rather than one request in isolation. It covers approvals, rate limits and running totals, ships under Apache 2.0, and is supported in AgentCore Policy, thou…