oscerd opened a new pull request, #1767: URL: https://github.com/apache/camel-website/pull/1767
A new blog post: **Authorizing what an AI agent may do in Apache Camel**. An AI agent decides which tools to call, and a tool can do real work. The model is not a security boundary, so the decision about what the agent may do has to sit on the tool call itself: SPIFFE for who is calling, Open Policy Agent for what they may do, evaluated in-process as WebAssembly (no OPA server, no network hop in the reasoning loop). The post covers the general pattern and walks through the new `ai-tools-spiffe-opa` example, now in camel-examples (https://github.com/apache/camel-examples/tree/main/ai-tools-spiffe-opa), including the prompt-injection case where a caller talked into asking for a refund is denied by the guard. It builds on the 4.22 "Camel Routes as AI Tools" post (`camel-ai-tool` and the MCP server) and the "Workload identity with SPIFFE and SPIRE" post. Written to publish under `oscerd`; content at `content/blog/2026/09/securing-ai-agent-tools/index.md`. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
