The Agent Skills Directory

[SAFE]: The skill provides documentation, code patterns, and protocol definitions for integrating LLM providers with the MolyKit framework. The code snippets follow standard Rust development practices for cross-platform applications.
[DATA_EXPOSURE]: A local file path (/Users/zhangalex/Work/Projects/FW/robius/moly/moly-kit) is present in the skill metadata. This is evaluated as a static documentation artifact from the author's development environment rather than a security threat or exfiltration attempt.
[INDIRECT_PROMPT_INJECTION]: The skill provides a surface for processing untrusted data via chat interfaces.
Ingestion points: Untrusted content enters the system context through the Message and MessageContent structures used by the Messages widget and BotClient trait in SKILL.md.
Boundary markers: No specific boundary markers or instruction delimiters are defined in the provided code templates.
Capability inventory: The skill includes tools for network requests to AI APIs (via BotClient::send) and asynchronous task spawning (via spawn).
Sanitization: No input sanitization or validation routines are included in the provided implementation examples.

molykit