The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to fetch documentation and model configuration from the vendor's official domain (pixazo.ai). These references are consistent with the skill's stated purpose and originate from the author's infrastructure.
[COMMAND_EXECUTION]: The skill provides integration examples using curl and system calls to chmod. The instructions specify setting the file permissions of the local credential file to chmod 600, which is a recommended security practice for protecting local secrets.
[INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect injection as it processes external API responses and user-provided URLs.
Ingestion points: The agent ingests data from user-provided image URLs and JSON responses from the gateway.pixazo.ai endpoint during the polling phase.
Boundary markers: The instructions do not specify explicit delimiters for the data returned from the external API.
Capability inventory: The skill uses network operations (POST/GET) to interact with the API and performs file system writes to store credentials in ~/.pixazo/api-key.
Sanitization: There are no specific instructions for sanitizing or validating the contents of the API response before presenting results to the user.

pixal3d