The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses tools like videos_analyze and videos_transcribe to process external video and audio content. This creates a surface for indirect prompt injection if the processed media contains malicious text or audio instructions intended to influence the agent's behavior.
Ingestion points: External video and audio files processed in SKILL.md and references/workflows.md.
Boundary markers: The instructions do not explicitly define delimiters for untrusted media content, though tool parameters like analysis_type provide some structure.
Capability inventory: The skill utilizes a specialized toolkit for video manipulation (generation, stitching, audio mixing) within the Hyper MCP environment.
Sanitization: No explicit sanitization of transcriptions or analysis results is mentioned before these are used in the agent's context.

video-generation