The Agent Skills Directory

[PROMPT_INJECTION]: The skill interpolates untrusted user inputs directly into the prompts used for image and video generation tools.
Ingestion points: The baby_description, baby_costume, and dialogue variables in SKILL.md are sourced from user input.
Boundary markers: There are no boundary markers or instructions to the model to ignore potential commands within the user-provided text.
Capability inventory: The skill has the capability to generate images and videos via the muapi CLI and perform network requests using curl to external endpoints.
Sanitization: No sanitization or validation logic is present to filter malicious instructions or escape content within the user inputs.
[COMMAND_EXECUTION]: The skill instructions include the use of the muapi CLI and curl for interacting with the MuAPI service. These commands are integral to the skill's intended functionality of generating media.
[DATA_EXFILTRATION]: The skill instructions describe using curl to send an API key (MUAPI_API_KEY) to https://api.muapi.ai. This operation targets the vendor's official API endpoint and is required for the skill to function.

muapi-talking-baby-video