The Agent Skills Directory

PROMPT_INJECTION (LOW): Potential for Indirect Prompt Injection. The skill ingests untrusted user text (the image prompt) and interpolates it into a Bash command.
Ingestion points: The --prompt and -p parameters used in the generate.py script as described in SKILL.md.
Boundary markers: Documentation suggests enclosing prompts in double quotes, which provides some protection but does not prevent escape sequences if the underlying script handles arguments unsafely.
Capability inventory: The skill is granted Bash and Read tool permissions, allowing it to execute local files and read outputs.
Sanitization: The implementation of scripts/generate.py is not provided in this analysis scope, preventing verification of argument sanitization.

invokeai-image-gen