The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface by interpolating untrusted user input directly into instructions for a video generation tool.
Ingestion points: User-supplied product names, features, and stories are used to replace [BRACKETED] placeholders in the prompt templates within SKILL.md.
Boundary markers: The instructions do not define clear delimiters or specific instructions for the agent to ignore potentially malicious commands embedded within user data.
Capability inventory: The skill is designed to trigger video generation via the rawugc-api skill using the POST /videos/generate endpoint.
Sanitization: There is no evidence of sanitization or validation instructions to handle malformed or adversarial user input before it is passed to the generation tool.

ugc-saas