The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the AI agent to execute shell commands using the muapi CLI and curl. These tools are used to configure authentication, generate UI mockups, and poll for task completion, which are core functions of the skill.
[EXTERNAL_DOWNLOADS]: The skill communicates with https://api.muapi.ai to send design requests and receive generated image assets. This is the official domain for the tool and service provider identified in the skill.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied content, including a url and redesign_style, and interpolates these values directly into prompts for image generation models. This creates a surface for indirect prompt injection, although the impact is restricted to affecting the visual output of the redesign process.
Ingestion points: The url and redesign_style inputs defined in SKILL.md are passed to generation prompts.
Boundary markers: There are no explicit markers or instructions used to separate user data from the system prompt or to warn the model to ignore instructions within the URL.
Capability inventory: The agent is granted shell command capabilities via muapi and curl to interact with the external API.
Sanitization: No input validation or sanitization routines are implemented for the user-provided URL or text parameters before they are processed.

muapi-url-to-design