The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the muapi CLI tool and curl to interact with an external API.
Evidence: The 'Notes for the Executing Agent' section explicitly directs the agent to run muapi auth configure, muapi image generate, and curl -X POST https://api.muapi.ai/api/v1/<endpoint>.
[INDIRECT_PROMPT_INJECTION]: The skill lacks sufficient boundaries and sanitization for user-provided data, creating a surface for indirect prompt injection when constructing prompts for the image generation models.
Ingestion points: User inputs defined in SKILL.md including brand_name, topic, style, color_scheme, and format.
Boundary markers: None. The skill uses simple variable interpolation (e.g., {{topic}}) directly inside prompt strings without delimiters or instructions to the model to ignore embedded commands.
Capability inventory: The skill possesses the ability to perform image generation, image editing, and network requests via curl and the muapi CLI.
Sanitization: None. There is no evidence of validation or escaping for the user-supplied strings before they are incorporated into the prompts sent to gpt4o-text-to-image or bytedance-seedream-v4.5.

muapi-brochures