The Agent Skills Directory

[SAFE]: The skill implements secure secret management by directing the agent to store the VIDEO_DB_API_KEY in a local .env file rather than hardcoding it or exposing it in the chat interface.
[SAFE]: External dependencies and network communication are restricted to official vendor resources, specifically the videodb Python package and the videodb.io API domain.
[INDIRECT_PROMPT_INJECTION]: The skill contains potential surfaces for indirect prompt injection as it processes untrusted data from media URLs and user-provided text.
Ingestion points: Media URLs in upload_media, text strings for titles/captions in render_timeline, and AI generative prompts in generate_music and generate_voice (SKILL.md).
Boundary markers: None provided to distinguish user data from instructions in the timeline composition.
Capability inventory: The skill has permissions to write to .env and execute Python code via the SDK path (SKILL.md).
Sanitization: No sanitization or validation of the provided URLs or text inputs is implemented.

video-edit