The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends installing external dependencies, including the belt-sh/cli tool and various model skill sets from the inference-sh GitHub organization, using the npx skills add command.
[INDIRECT_PROMPT_INJECTION]: The skill accepts untrusted external data via its core parameters (prompt, images, mask) which are then used as input for the belt CLI tool.
Ingestion points: prompt, images, and mask parameters within the usage examples in SKILL.md.
Boundary markers: None observed in the skill instructions to separate user input from commands or instructions.
Capability inventory: The skill utilizes the Bash tool to interact with the local system and execute the belt CLI application.
Sanitization: There is no evidence of input validation, URL filtering, or prompt sanitization prior to the execution of the image generation commands.

gpt-image