The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill manages authentication by reading from and writing to ~/.pixazo/api-key. It instructs the agent to set file permissions to chmod 600, which is consistent with standard secret management practices for developer tools.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to gateway.pixazo.ai to submit video processing tasks and retrieves updated documentation or instruction schemas from www.pixazo.ai. These resources are owned by the skill vendor.
[COMMAND_EXECUTION]: The skill includes instructions to use npx to manage other skills within the same vendor ecosystem and utilizes chmod to secure local credential files.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest external instructions and parameter schemas from a remote URL to guide agent behavior during API calls.
Ingestion points: Fetches markdown instructions from https://www.pixazo.ai/models/video-to-previs.md to determine valid parameters and polling logic.
Boundary markers: None identified.
Capability inventory: File system writes (~/.pixazo/api-key) and network operations (cURL, Python requests, and Node.js fetch).
Sanitization: None identified.

video-to-previs