DevOps / SRE / Platform · 06.08.2026, 19:23 UTC
Your AI agent’s next tool call may be valid but wrong. AWS’s Dogwood promises to fix that.
| Schweregrad | info |
|---|---|
| Kategorie | DevOps / SRE / Platform |
| Quelle | The New Stack ↗ |
| Veröffentlicht | 06.08.2026 UTC |
Sicherheitsmeldung mit Schweregrad noch nicht bewertet. Technische Details im Tab „Originaltext“; empfohlene Schritte in der Checkliste.
AWS on Thursday launched Dogwood, an open-source policy language and reference interpreter that lets developers govern sequences of AI agent tool calls instead of evaluating each action in isolation.
The company has also added Dogwood support to Amazon Bedrock AgentCore Policy, its managed service for controlling which tools an agent may call and under what conditions. The language and its reference implementation are now available under the Apache 2.0 license.
Dogwood builds on Cedar, the open source authorization language that already powers AgentCore Policy. Cedar, which AWS contributed to the CNCF as a sandbox project in late 2025, can decide whether a user or agent may call a refund tool with a given set of arguments, for example. Dogwood takes this a step further because it can also take earlier events into account. That may be whether somebody approved the refund, how much the agent has refunded in the past hour, or whether it previously accessed information that should prevent it from contacting an outside service.
“Point-in-time decisions make sense for many forms of access control, but when agents compose multiple actions into longer workflows, the sequence itself becomes something teams want to govern.”
As the AWS team explains, “Point-in-time decisions make sense for many forms of access control, but when agents compose multiple actions into longer workflows, the sequence itself becomes something teams want to govern. Dogwood gives them a language for expressing policies over sequences, in order to capture constraints on prerequisites, rate limits, and …