The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill takes user-provided strings for baby descriptions, costumes, and dialogue and interpolates them directly into prompts for image and video generation models. This creates a surface where malicious input could influence the behavior of the downstream AI models.
Ingestion points: SKILL.md via the {{baby_description}}, {{baby_costume}}, and {{dialogue}} placeholders.
Boundary markers: None present to delimit user content or instruct the model to ignore embedded instructions.
Capability inventory: The skill uses muapi CLI commands and curl to perform network operations and generate media.
Sanitization: No evidence of input validation, escaping, or filtering of external content.
[COMMAND_EXECUTION]: The skill instructions direct the AI agent to execute shell commands using the muapi CLI tool. It also provides a fallback mechanism using curl to interact with the api.muapi.ai endpoint, which is the legitimate infrastructure for the skill's stated purpose.

muapi-talking-baby-video