dlsps-user

Fail

Audited by Gen Agent Trust Hub on Aug 19, 2026

Risk Level: HIGHCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill instructs the user to execute sudo ufw disable in references/troubleshooting.md to resolve WebRTC connectivity issues. Disabling the system firewall is a high-risk operation that compromises host security by removing network access controls.\n- [COMMAND_EXECUTION]: The skill relies on shell command execution for its primary purpose, including docker compose, helm install, and system-level environment variable configuration (RENDER_GID via stat). While necessary for the service, these represent a high-privilege interaction surface with the host system.\n- [EXTERNAL_DOWNLOADS]: The skill performs external downloads from well-known sources including the Intel Docker Hub repository and the open-edge-platform GitHub repository. These operations are documented for service deployment and target established vendor infrastructure.\n- [INDIRECT_PROMPT_INJECTION]: The skill exposes an attack surface by ingesting untrusted data through a REST API (e.g., video URIs and pipeline parameters) without documented validation.\n
  • Ingestion points: REST API POST requests to /pipelines/{name}/{version} (described in SKILL.md).\n
  • Boundary markers: None identified.\n
  • Capability inventory: Execution of GStreamer pipelines with hardware acceleration and network publishing capabilities.\n
  • Sanitization: No specific sanitization or validation of input URIs or parameters is mentioned.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Aug 19, 2026, 08:34 PM
Security Audit — agent-trust-hub — dlsps-user