The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill manages a Pixazo API key by instructing the agent to check environment variables, a local file, or the current conversation. It prompts the user to store the key in ~/.pixazo/api-key and explicitly mandates the use of chmod 600 to secure the file. This represents standard and secure credential management for CLI-based agents.
[EXTERNAL_DOWNLOADS]: The skill fetches additional model reference documentation and status polling logic from the vendor's official website at pixazo.ai. These downloads originate from a trusted source associated with the skill author.
[COMMAND_EXECUTION]: The skill utilizes common utilities like curl for API requests and chmod for securing configuration files. It also provides a command to install related skills from the vendor's repository using npx.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a standard vulnerability surface for indirect prompt injection as it processes JSON data returned from the Pixazo API gateway.
Ingestion points: Polling responses from the status endpoint at gateway.pixazo.ai (SKILL.md).
Boundary markers: None present.
Capability inventory: File system access (~/.pixazo/api-key), network operations (curl, requests, fetch), and shell command execution (chmod).
Sanitization: None described for the API status responses.

kandinsky