The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied image URLs and text descriptions to generate visualizations. While this constitutes an attack surface where malicious instructions could be embedded in metadata or image content, it is the primary intended function of the skill.
Ingestion points: User-provided image_urls and message content in SKILL.md examples.
Boundary markers: The skill relies on the EachLabs API backend to handle content safety and boundary enforcement.
Capability inventory: The skill is restricted to making HTTPS requests to the EachLabs API endpoint.
Sanitization: Input validation and safety filtering are performed by the EachLabs core agent service.
[COMMAND_EXECUTION]: The documentation includes illustrative curl commands. These are standard integration examples for developers to interact with the EachLabs API and do not perform any unauthorized or dangerous system operations.
[DATA_EXPOSURE]: The skill correctly demonstrates the use of an environment variable ($EACHLABS_API_KEY) for authentication rather than hardcoding sensitive credentials. Network operations are directed to the vendor's own API domain (eachlabs.run), representing normal functionality.

interior-design-visualization