The Agent Skills Directory

[PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection where untrusted user inputs are interpolated directly into prompts for image and video generation tools.
Ingestion points: User-provided inputs including product_name, campaign_message, target_audience, and visual_style are ingested via SKILL.md.
Boundary markers: The prompt templates do not utilize delimiters (e.g., XML tags or triple quotes) or negative constraints to prevent the model from following instructions embedded within the user data.
Capability inventory: The skill has the ability to execute shell commands via the muapi CLI and perform network operations via curl.
Sanitization: No input validation, escaping, or filtering is applied to the user-supplied text before interpolation.
[COMMAND_EXECUTION]: The skill relies on the execution of shell commands to interact with the MuAPI platform.
Evidence: Instructions specify the use of muapi image generate, muapi image edit, and muapi video generate CLI tools.
[DATA_EXFILTRATION]: The skill performs network operations to transmit data to an external service endpoint.
Evidence: The skill communicates with https://api.muapi.ai/api/v1/ using the MUAPI_API_KEY to process campaign assets. This is the official endpoint associated with the skill's stated purpose.

muapi-product-campaign