Executive brief
A security vulnerability exists in rt-claw, a software component used for AI assistant communication. An attacker can remotely force the software to make unauthorized web requests to internal systems or local services that are normally protected. This could allow an attacker to steal sensitive internal data or gain a foothold in a private network.
Technical details
A Server-Side Request Forgery (SSRF) vulnerability exists in the swarm RPC service of zevorn rt-claw up to version 0.2.0. The 'receiver_thread' function in 'claw/services/swarm/swarm.c' dispatches unauthenticated UDP RPC packets to 'handle_rpc_request', which invokes the 'http_request' tool with attacker-controlled JSON parameters. On Linux builds, the 'url' parameter is passed directly to libcurl via 'claw_net_get' without validation of the destination address. A remote attacker can exploit this to coerce the application into fetching resources from loopback (127.0.0.1) or internal network endpoints and returning the response body in the RPC reply. As of the advisory date, the project has not responded to the issue report.
Affected products
- zevorn rt-claw up to 0.2.0
Timeline
- 2026-06-12: disclosed: Vulnerability discovered and verified via PoC
- 2026-07-18: advisory: CVE-2026-16128 published via VulDB/NVD