The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data through the user-provided person_image URL. The agent is instructed to analyze the setting, mood, and activity within this image to generate relevant stickers and captions, creating a surface where malicious content embedded in the image could influence the model's behavior.
Ingestion points: The person_image input parameter defined in SKILL.md.
Boundary markers: No explicit markers or instructions are provided to the agent to treat the image content as untrusted data.
Capability inventory: The skill utilizes the muapi CLI and curl for network-based image generation and API interactions.
Sanitization: There is no evidence of filtering or validation for the content of the provided image URL.
[COMMAND_EXECUTION]: The skill requires the agent to execute shell commands for service configuration and image editing. This includes using the muapi CLI tool (muapi auth configure, muapi image edit) and curl for direct API interaction with the api.muapi.ai infrastructure.

muapi-chibi-collage-effect