Retrieve decisions, documentation, learnings, and patterns from a RAG corpus with filters.
npx skills add https://github.com/arbgjr/smart_alarm --skill rag-query
Please help me install this Agent Skill. Skill: rag-query Source: https://github.com/arbgjr/smart_alarm/tree/main/.claude/skills/rag-query Command: npx skills add https://github.com/arbgjr/smart_alarm --skill rag-query
Esta skill facilita a consulta ao corpus de conhecimento do projeto (RAG), permitindo localizar decisões, documentação, aprendizados e padrões armazenados, para embasar decisões e reduzir buscas manuais.
Use rag-query to search for a topic like "autenticação" and retrieve related decisions, docs, patterns, and learnings from the project knowledge base.
High-intent search queries and answers about installing and using this skill.
You can retrieve ADRs, engineering patterns, and learnings from your project corpus by executing keyword-based searches with optional phase and concept filters. This mechanism uses graph-based expansion and decay boosting to return context-aware metadata and relevance scores.
Yes, you can filter semantic search results by applying optional phase and concept constraints alongside your keyword queries. This refines the retrieval of decisions, docs, and patterns to provide context-aware results specifically matched to your current project phase.
The RAG retrieval mechanism works by combining a lightweight text index with a graph-based expansion module and a simple decay boosting mechanism. This filters and ranks your project's decisions and learnings to return metadata, provenance, and relevance scores.
No, you do not need a specific database dependency to query project learnings and patterns. The query system relies on a lightweight text index and requires only the pyyaml dependency to parse configurations and retrieve your stored knowledge corpus.
You should use targeted retrieval instead of manual searches because it quickly locates known engineering patterns and learnings to prevent repeated errors. It retrieves metadata, provenance, and relevance scores to support continuous improvement and decision-making.
sambeauBuild and query a codebase knowledge-graph to reveal structure and dependencies.
semicolon-devteamDetect and prioritize technical debt in TypeScript codebases.
simonheimlicherCreate Architecture Decision Records for TypeScript architecture decisions.
dalagerEnable high-speed semantic vector search with HNSW indexing and metadata filtering.
