Skills
Navigate
skills/agent-media-ai/agent-media/publish-to-social/Gen Agent Trust Hub

publish-to-social

Pass

Audited by Gen Agent Trust Hub on Oct 16, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data, such as video captions and URLs, representing a standard attack surface for indirect prompt injection. A malicious input in the video metadata or caption could attempt to influence the agent's behavior.
  • Ingestion points: The caption and video_url parameters in the social_publish tool and API endpoints.
  • Boundary markers: No specific delimiters or safety instructions are defined in the skill instructions to separate data from commands.
  • Capability inventory: The skill allows network operations (publishing via REST API) and command-line execution through the agent-media CLI.
  • Sanitization: The skill does not explicitly mention client-side sanitization of the caption or URL, though validation is expected to occur at the API level.
  • [EXTERNAL_DOWNLOADS]: The skill interacts with api.agent-media.ai and uses video_url from R2 storage (pub-...r2.dev). These are official vendor resources and necessary for the skill's primary function of media publishing.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 16, 2026, 04:41 PM