The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides structured curl command templates intended for execution to interact with the EachLabs API. These templates include instructions to use environment variables for authentication and specific flags to modify API behavior, such as disabling safety filters.
[INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by ingesting untrusted data from users and external sources which is then passed to a remote generation engine.
Ingestion points: The skill accepts natural language strings in the messages parameter and remote image references via the image_urls array defined in SKILL.md.
Boundary markers: None provided; the instructions do not implement delimiters or guardrail prompts to prevent the agent from following instructions potentially embedded within user-supplied content or image metadata.
Capability inventory: The skill performs network operations using curl to transmit data to the eachsense-agent.core.eachlabs.run endpoint.
Sanitization: There is no evidence of input validation, filtering, or escaping for the user-provided descriptions or external image URLs before they are processed by the skill.

nsfw-content-generation