The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data, such as text prompts and image URLs, which are then interpolated into network requests and image processing operations.
Ingestion points: User-supplied strings for prompts and reference images are used to construct API calls in SKILL.md.
Boundary markers: Absent; there are no specific delimiters or instructions to the agent to treat user-provided data as potentially containing instructions.
Capability inventory: The skill utilizes curl, Python's requests library, and the Node.js fetch API for external network communication. It also performs local file writes.
Sanitization: No explicit validation or escaping of the user-provided prompt or URL content is defined.
[COMMAND_EXECUTION]: The skill directs the agent to perform file system operations, including writing a configuration file and setting its permissions.
Evidence: Step 1 of SKILL.md instructs the agent to save the API key to ~/.pixazo/api-key and apply chmod 600 to secure the file.

codeformer