The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill accepts user-supplied text for the 'subject_description' which is interpolated directly into prompts for image generation models without sanitization.
Ingestion points: The subject_description input in SKILL.md is the primary entry point for untrusted data.
Boundary markers: No specific delimiters or safety instructions are used to separate user input from the rest of the prompt template.
Capability inventory: The skill utilizes muapi CLI tools and curl for network requests to generate and edit images.
Sanitization: There is no evidence of input validation or filtering for potentially malicious instructions hidden within the subject description.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands, including muapi CLI tools and curl, to interact with the vendor's API.
The commands include muapi image generate, muapi image edit, and raw curl requests to https://api.muapi.ai.
While these are core functionalities of the skill, they represent a capability surface that could be targeted by prompt injection.

muapi-multi-angle-reshoot