The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were identified in the analysis. The skill correctly implements the iii engine architecture for registering workers and handlers.
[COMMAND_EXECUTION]: The skill defines the trigger() primitive used for cross-language function invocation as a core framework feature for worker communication.
[DATA_EXFILTRATION]: Mentions connection to a local engine instance via ws://localhost:49134. This is a standard development configuration for local testing and does not pose an exfiltration risk.
[PROMPT_INJECTION]: The documentation describes an architectural surface for processing external data from triggers (HTTP, queues, streams). This is documented neutrally as a standard framework capability rather than a malicious injection pattern.

iii-functions-and-triggers