The Agent Skills Directory

[SAFE]: The skill serves as a technical reference and implementation guide for LiteLLM-RS A2A Protocol. It contains Rust code snippets for building a JSON-RPC 2.0 gateway to orchestrate tasks across different AI agent providers.
[DATA_EXPOSURE]: The implementation defines structures for handling authentication (API keys, Bearer tokens, OAuth2). The provided configuration examples correctly use environment variable placeholders (e.g., ${LANGGRAPH_API_KEY}) rather than hardcoding credentials.
[COMMAND_EXECUTION]: The code implements network operations using the reqwest library to communicate with provider endpoints (such as Google Vertex AI). These operations are part of the core functionality of an agent gateway and target well-known cloud services.
[INDIRECT_PROMPT_INJECTION]: The protocol defines how natural language instructions are passed between agents via the TaskInput structure. While this represents a standard attack surface for LLM-based systems, the skill is a protocol definition and does not contain any malicious injection patterns or unsafe prompt construction.

a2a-protocol