The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied text for speech synthesis, which could potentially contain malicious instructions intended for the agent or the underlying model.
Ingestion points: The text parameter in the API request body within SKILL.md.
Boundary markers: Absent; user text is interpolated directly into the JSON request body.
Capability inventory: File system read/write (managing ~/.pixazo/api-key), network requests via curl, requests, and fetch to external API endpoints.
Sanitization: Absent; no specific escaping or validation of the input text is mentioned.
[EXTERNAL_DOWNLOADS]: The skill fetches model schemas and detailed documentation from the vendor's official domain at https://www.pixazo.ai/models/voxcpm.md to assist the agent in configuring API calls.
[COMMAND_EXECUTION]: The skill instructs the agent to perform security-sensitive operations on the local system.
It uses chmod 600 to secure the ~/.pixazo/api-key file, which is a recommended security practice for protecting secrets.
It provides an example npx command for installing additional vendor-supplied skills.

voxcpm