Analyzes RAG pipeline retrieval quality by computing embedding similarities, detecting outlier chunks, and recommending thresholds. Use when: debug rag retrieval, analyze embedding similarity, tune similarity threshold, compare embedding models, evaluate retrieval quality.
rag-retrieval-analyzer analyzes RAG pipeline retrieval quality by computing embedding similarities, detecting outlier chunks, and recommending thresholds. Use when: debug rag retrieval, analyze embedding similarity, tune similarity threshold, compare embedding models, evaluate retrieval quality.
See the source code for full details and usage instructions.
Copy the skill into your OpenClaw skills directory:
git clone https://github.com/NeoSkillFactory/rag-retrieval-analyzer.git
cp -r rag-retrieval-analyzer ~/.openclaw/skills/rag-retrieval-analyzerMIT © NeoSkillFactory