The Agent Skills Directory

[DATA_EXFILTRATION]: Transfers image URLs and user prompts to the vendor's official API endpoint (sense.eachlabs.run) for processing. This network activity is restricted to the skill's primary function and targets verified infrastructure.
[PROMPT_INJECTION]: Ingests data through image_urls and message fields as seen in SKILL.md. While explicit boundary markers and sanitization methods are not specified, the skill documented in references/SSE-EVENTS.md provides capabilities for execute_model and web_search_query as part of its standard workflow.
[COMMAND_EXECUTION]: Includes documentation for curl commands to interact with the EachLabs API. These examples demonstrate secure practices by using environment variables ($EACHLABS_API_KEY) for authentication rather than hardcoded secrets.

style-transfer