The Agent Skills Directory

[SAFE]: The skill correctly uses an environment variable ($EACHLABS_API_KEY) for API authentication, avoiding hardcoded secrets and following best practices for credential management.
[SAFE]: All network operations are performed against official vendor-owned domains (eachsense-agent.core.eachlabs.run, storage.eachlabs.ai, and eachlabs.ai) for the purpose of communicating with the each::sense AI service.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted user input via the messages array in API requests sent to the remote AI service. This constitutes a standard injection surface; however, the risk is mitigated as processing is handled by the vendor's remote infrastructure which is the primary intended function of the skill.
Ingestion points: User-provided content in the messages object within SKILL.md examples.
Boundary markers: None present in the prompt examples.
Capability inventory: Network POST requests via curl to eachlabs.run endpoints.
Sanitization: Not present in the client-side skill description; assumed to be handled by the remote AI endpoint.

seamless-pattern-generation