The Agent Skills Directory

[SAFE]: The skill provides standard documentation and code examples for Unirest-Java 4.x, which is a legitimate and widely used open-source networking library.\n- [EXTERNAL_DOWNLOADS]: The documentation references Maven and Gradle dependencies from com.konghq (Kong). These are well-known and official artifacts for the Unirest library, and their use in this context is appropriate and safe.\n- [DATA_EXFILTRATION]: Code examples contain generic placeholders for Basic Authentication and Proxy credentials (e.g., "user", "pass"). These are standard examples for demonstrating library features and do not indicate any attempt to harvest or hardcode real credentials.\n- [COMMAND_EXECUTION]: The skill focuses entirely on Java API calls for networking and object mapping. No shell commands, subprocess calls, or other patterns associated with unauthorized command execution were found.\n- [INDIRECT_PROMPT_INJECTION]: The skill documents the library's ability to fetch data from external sources (APIs, SSE, WebSockets), which is its primary purpose. While this fetches external content into the application context, the skill does not suggest using this content in prompts or other sensitive operations.\n
Ingestion points: Data enters through Unirest.get(), Unirest.sse(), and Unirest.webSocket() calls documented in references/requests.md, references/sse.md, and references/websocket.md.\n
Boundary markers: The documentation focuses on standard library usage and does not explicitly discuss prompt engineering or security boundaries for the fetched data.\n
Capability inventory: The library provides network access and file writing capabilities (via asFile()).\n
Sanitization: The skill covers standard JSON and object mapping but does not involve interpolation of external data into LLM prompts.

unirest-java-4