Query multiple registered knowledge graphs with kind-based filtering for code, docs, and metadata.
npx skills add https://github.com/Flux-Frontiers/doc_kg --skill kgrag-usage
Please help me install this Agent Skill. Skill: kgrag-usage Source: https://github.com/Flux-Frontiers/doc_kg/tree/main/.claude/skills/kgrag-usage Command: npx skills add https://github.com/Flux-Frontiers/doc_kg --skill kgrag-usage
KGRAG solves the problem of searching and extracting useful snippets across multiple knowledge graphs consistently, even when they span different repositories and KG kinds.
Use KGRAG to search for implementations by running kgrag query "API authentication" --kind code.
High-intent search queries and answers about installing and using this skill.
Federated retrieval across multiple registered knowledge graphs enables simultaneous code and documentation snippet extraction by querying a central KG registry with kind-based filtering to return deterministic packaged results.
Multi-KG snippet packing collects graph-grounded passage packs and source-context snippets across different repositories, filtering by kind (code, doc, meta) to produce deterministic packaged results suitable for LLM context.
Automated initialization detects repository layers to build the index, then registers the new knowledge graph in the central KG registry so cross-KG queries can access the dataset alongside existing instances.
Kind-based filtering allows queries to target specific content kinds (code, doc, meta), so users can isolate relevant implementations, documentation, or conceptual descriptions across all registered CodeKG, DocKG, and MetaKG instances.
Cross-repo pattern discovery requires querying an aggregated registry state to ensure deterministic packaged results; inconsistent outputs often indicate the central KG registry needs a health audit to verify queries run against correct indexed datasets.
Flux-FrontiersQuery multiple knowledge graphs via a unified registry and CLI.
braveRetrieve pre-extracted web content with text, tables, and code for LLM grounding.
oakossPackage codebases into compressed context bundles using Repomix, Gitingest, and llm-tldr.
handsomelong922Coordinates AI search queries across platforms via Chrome DevTools sessions.
