skills/zoom/skills/meeting-sdk/linux/Gen Agent Trust Hub

meeting-sdk/linux

Pass

Audited by Gen Agent Trust Hub on Aug 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The documentation provides C++ code examples that use the system() function to call FFmpeg for merging raw audio and video files.
  • [PRIVILEGE_ESCALATION]: Setup instructions frequently require the use of sudo for installing system dependencies and manually managing packages with dpkg.
  • [INDIRECT_PROMPT_INJECTION]: The skill outlines architectures that ingest live meeting content for analysis by external AI services (e.g., Anthropic, AssemblyAI).
  • Ingestion points: Raw PCM audio and YUV video streams, along with meeting chat messages.
  • Boundary markers: None are implemented in the provided implementation snippets.
  • Capability inventory: The bot can capture raw media, join or leave meetings, and interact with the in-meeting chat system.
  • Sanitization: No sanitization of meeting content is performed before it is passed to AI models.
  • [EXTERNAL_DOWNLOADS]: The skill directs users to fetch the Meeting SDK and sample repositories from official Zoom domains and GitHub. It also includes instructions for downloading platform-specific packages from official Ubuntu security mirrors.
  • [REMOTE_CODE_EXECUTION]: The documentation involves downloading and installing Debian packages (.deb) for OpenSSL 1.1 from official Ubuntu security mirrors using dpkg.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 26, 2026, 09:56 PM
Security Audit — agent-trust-hub — meeting-sdk/linux