Tag
Cross-Origin Message Handling Vulnerability in Open WebUI
4 TTPs 1 CVEOpen WebUI versions 0.7.0 through 0.11.3 contain a cross-origin message handling vulnerability that allows unauthenticated websites to steal a user's session JWT via a crafted postMessage payload.
Joro: Unauthenticated Cross-Origin Plugin Upload Leads to RCE
2 rules 5 TTPs 1 IOCJoro's default proxy mode (versions ≤ v1.1.0) is vulnerable to unauthenticated remote code execution (CVE-2026-53649) via a local API on `127.0.0.1:9090` that allows cross-origin JavaScript to upload a malicious native plugin and trigger a system restart, leading to RCE as the operator's user from a single page visit.
Network-AI Unauthenticated Cross-Origin MCP Tool Invocation via Empty Default Secret (CVE-2026-46701)
2 rules 1 TTPNetwork-AI is vulnerable to an unauthenticated cross-origin attack due to an empty default secret and permissive CORS configuration, allowing an attacker to lure a user to a malicious web page and invoke MCP tools like config_set, agent_spawn, and blackboard_write against a default-configured localhost server.
cpp-httplib Vulnerability Leads to Credential Leakage via HTTP Redirects
2 rules 1 TTPThe cpp-httplib library prior to version 0.39.0 forwards stored authentication credentials to arbitrary hosts via HTTP redirects, potentially exposing sensitive information to malicious actors.
OpenClaw fetchWithSsrFGuard replays unsafe request bodies across cross-origin redirects
2 rules 1 TTPThe `fetchWithSsrFGuard` function in OpenClaw replays unsafe request bodies across cross-origin redirects, potentially leading to the exposure of sensitive information when following cross-origin redirects in versions prior to 2026.4.8.
Cline Kanban Server Cross-Origin WebSocket Hijacking Vulnerability
3 rules 4 TTPs 1 IOCThe `kanban` npm package, used by the `cline` CLI, has a cross-origin WebSocket hijacking vulnerability. Due to the lack of Origin header validation, any website can connect to the kanban server via WebSocket and leak sensitive data, hijack running AI agent terminals leading to remote code execution, or kill running agent tasks, resulting in information disclosure, RCE, and denial of service.