Telos

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 0.90). The skill includes an explicit, "MANDATORY" instruction to immediately execute a local HTTP curl notification (http://localhost:31337/notify) before any action, a forced external side-effect unrelated to the described file-analysis/update functionality that could be used for exfiltration or to change runtime behavior.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). This codebase exposes and reads a user's private TELOS files via unauthenticated HTTP endpoints, runs a local inference script with the full personal context (creating an easy remote-execution/exfiltration vector), and documents shipped credentials — together these create high-risk data-exfiltration and remote-code-execution/backdoor capabilities.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The runtime workflow path DashboardTemplate/app/api/chat/route.ts reads outsider-authored free text from the HTTP POST body (message) and sends it—along with first-party TELOS file contents from getTelosContext()—into Inference.ts via spawn, creating an indirect prompt-injection surface.