The Agent Skills Directory

[SAFE]: The skill serves as an educational and structural reference for building RAG pipelines. All code patterns follow industry standards for performance and reliability.
[EXTERNAL_DOWNLOADS]: The skill templates reference reputable APIs and services including OpenAI, Anthropic, Voyage AI, Cohere, and Tavily. These are standard external resources for the documented functionality.
[DATA_EXFILTRATION]: There are no indicators of unauthorized data harvesting. The skill encourages best practices for security, such as utilizing environment variables for API credentials rather than hardcoding them into scripts.
[PROMPT_INJECTION]: The skill provides defensive system prompt templates designed to prevent hallucinations and model bypass by grounding responses strictly in provided context, which is a key security feature in RAG systems.
[COMMAND_EXECUTION]: The implementation examples use standard libraries for database operations and network requests. No patterns for arbitrary shell command execution or unsafe system calls were identified.

rag-retrieval