twelvelabs-video-understanding
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of informational documentation for interacting with the TwelveLabs API. It proactively addresses security and compliance concerns by including specific guidance on biometric data privacy, the requirement for user consent under laws like GDPR and BIPA, and strategies for validating AI output accuracy.\n- [EXTERNAL_DOWNLOADS]: The skill references official TwelveLabs API endpoints (api.twelvelabs.io) and their legitimate Python SDK on GitHub. These references are standard for the skill's intended purpose and target a well-known technology service.\n- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for processing external video data via user-supplied URLs, which represents a potential attack surface for indirect prompt injection.\n
- Ingestion points: User-provided video files and public URLs for indexing and analysis tasks in SKILL.md.\n
- Boundary markers: The skill includes extensive privacy warnings and instructions for agents to verify user rights and consent before processing identifiable individuals.\n
- Capability inventory: Performs network API requests to TwelveLabs for analysis and search; results are processed by the agent.\n
- Sanitization: Recommends human-in-the-loop verification, low temperature settings for factual accuracy, and anchoring claims to verbatim on-screen evidence.
Audit Metadata