Configure hybrid search weights and minimum score thresholds via CLI.
npx skills add https://github.com/alsk1992/CloddsBot --skill search-config
Please help me install this Agent Skill. Skill: search-config Source: https://github.com/alsk1992/CloddsBot/tree/main/src/skills/bundled/search-config Command: npx skills add https://github.com/alsk1992/CloddsBot --skill search-config
This Skill allows you to precisely control how search queries are processed, balancing keyword matching with semantic understanding to retrieve the most relevant information.
Show the current search configuration and weights.
High-intent search queries and answers about installing and using this skill.
Tune hybrid search parameters by adjusting the weights for keyword and semantic matching. You can balance BM25 and vector search weights to prioritize exact term matches or conceptual similarity, and set minimum score thresholds to filter retrieval results.
Increase BM25 weight when you need exact keyword matching, such as searching for specific technical documentation or precise function names. Increase vector weight when queries require conceptual understanding and semantic similarity rather than exact term retrieval.
Yes, you can manage search configuration via CLI commands. The system allows you to view current parameter settings, set new weights for BM25 and vector retrieval, and reset configurations to their default state.
The best way to export search configuration details is using the built-in export function. This feature outputs your current hybrid search parameters, including vector and BM25 weights and minimum score thresholds, for analysis and deployment optimization.
Set minimum score thresholds through the configuration management interface to filter low-relevance search results. This ensures that only documents meeting a specific relevance score from the hybrid retrieval process are returned to the user.
Yes, hybrid search tuning operates without external dependencies. You can configure vector and BM25 weights, adjust minimum score thresholds, and export settings natively through the CLI without installing additional packages or modules.
steipeteIndex local files and search them with BM25, vector, and hybrid queries.
sourcegraphCompute precision, recall, MRR, nDCG, and MAP for file retrieval comparisons.
vasilyu1983Design and implement RAG retrieval systems with hybrid search and reranking.
Sir-chawakornDesign Elasticsearch mappings, HNSW parameters, and hybrid RRF ranking strategies.
