The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation includes examples using curl to interact with eachsense-agent.core.eachlabs.run. This is the official API endpoint for the skill's author, eachlabs, and is a standard implementation for this type of service.
[CREDENTIALS_UNSAFE]: The skill demonstrates safe secret management by using the $EACHLABS_API_KEY environment variable placeholder in its examples, rather than hardcoding sensitive credentials.
[INDIRECT_PROMPT_INJECTION]: The skill defines a surface for processing external data via the image_urls parameter. This is the intended primary purpose of the image outpainting functionality. The risk is minimized as the agent is directed to a specific, narrow task (image extension) through the vendor's API.
[DATA_EXFILTRATION]: Network operations are directed towards official vendor domains (eachlabs.run, eachlabs.ai) for functional purposes. No unauthorized data transmission or access to sensitive local files was detected.

image-outpainting