videodb
Pass
Audited by Gen Agent Trust Hub on Aug 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is a standard integration for the VideoDB media platform. It focuses on media ingestion, indexing, and editing without any hidden or malicious behaviors.- [COMMAND_EXECUTION]: To support real-time capture and SDK interactions, the skill uses a local Python script (scripts/ws_listener.py) and inline Python code. The listener script follows good security hygiene by ensuring that its state directory is created with private permissions (0o700) to protect session tokens and event logs.- [INDIRECT_PROMPT_INJECTION]: The skill enables the analysis of video transcripts and visual scenes using LLMs. While this processes external data, it is a core feature for video understanding.
- Ingestion points: Media transcripts and AI-generated scene descriptions extracted from media files as described in SKILL.md and generative.md.
- Boundary markers: None explicitly shown in example prompt templates; data is directly concatenated.
- Capability inventory: SDK access, shell script execution for capturing sessions, and Python script execution.
- Sanitization: The skill relies on the service's AI models to process the data safely rather than implementing custom sanitization logic.
Audit Metadata