Sentinel controls AI agent tool use without extra token cost
Sentinel is an SDK for controlling what an AI agent is allowed to do while it runs. It checks a tool call before the tool is executed, instead of only reporting problems afterward.
The rules can cover role boundaries, intended action, and hard blocks that stop specific behavior. The monitoring path does not call another model, so the checks are and do not add .
A Claude Code adapter is already available, and installation is through an . The developer is asking which should be exposed next.
Key points
- Sentinel checks an AI agent’s tool call before the tool runs.
- Rules can cover role boundaries, intent, and hard block conditions.
- It avoids extra , so it adds no .
- A Claude Code adapter is live.
- The project is and available as an .