The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input via the floor_plan_description parameter and interpolates it directly into prompts for image generation tools and API payloads.
Ingestion points: The floor_plan_description input defined in SKILL.md.
Boundary markers: The instructions lack delimiters or boundary markers to isolate user input from the rest of the prompt, increasing the risk that embedded instructions could be obeyed by the underlying model.
Capability inventory: The skill has the ability to execute CLI commands (muapi) and make network requests (curl).
Sanitization: There is no evidence of sanitization, validation, or escaping of the user-provided text before interpolation.
[COMMAND_EXECUTION]: The skill's fallback instructions direct the agent to construct curl commands by substituting {{input_name}} placeholders with user-provided text. This practice creates a command injection risk if the executing agent does not properly escape shell metacharacters present in the user input.
[DATA_EXFILTRATION]: The skill facilitates the transmission of user-provided descriptions and the MUAPI_API_KEY credential to an external endpoint (api.muapi.ai). While this is necessary for the skill's functionality, the destination domain is not part of a standard trusted whitelist.

muapi-floor-plan-rendering