The Agent Skills Directory

[SAFE]: The skill is entirely informational and does not include any executable code, scripts, or tool definitions. It provides architectural best practices for RAG pipelines, including chunking, embedding selection, and evaluation strategies.
[NO_CODE]: The skill contains no code files or referenced scripts, posing no risk of remote code execution or privilege escalation.
[SAFE]: No prompt injection patterns were detected. The instructional language is directed at teaching RAG concepts and does not attempt to override agent safety filters or system prompts.
[SAFE]: No data exfiltration or credential exposure patterns were identified. Technical terms and examples are used purely for illustrative purposes within a search context.

rag-architecture