The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to the way it handles user-provided data.
Ingestion points: User-provided content and topics are ingested directly into the generate_image prompt template as seen in SKILL.md.
Boundary markers: The skill lacks any delimiters (e.g., triple quotes or xml tags) or instructions to ignore embedded commands within the user content.
Capability inventory: The skill has the capability to call generate_image, edit_image, and continue_editing based on the prompt instructions in SKILL.md.
Sanitization: No validation or sanitization is performed on the user-supplied strings before they are interpolated into the tool prompt.

infographic-old