Context·Context
Repository intelligence and knowledge graph tools
Map codebases as graphs—dependencies, symbols, impact—so agents retrieve structure, not only embeddings.
This article is AI-assisted and co-authored by Xesca Alabart, co-founder of EasySpecs.
Repository intelligence tools model relationships—imports, calls, ownership, impact—so agents (and humans) navigate large systems without reading every file. This is context engineering at the structural layer: complementary to RAG over text chunks.
Tools in this category
| Tool | Notes |
|---|---|
| Augment Context Engine | Real-time KG, dependency paths; MCP exposure. |
| GitLab Knowledge Graph | Repo → graph via MCP in Duo Agent Platform. |
| Bevel Code-to-Knowledge-Graph | JVM/LSP, Neo4j, impact analysis. |
| FalkorDB Code Graph | Graph DB + NL queries. |
| GraphGen4Code | IBM research toolkit. |
| Aider Repository Map | tree-sitter + PageRank; dynamic token budget. |
| Understand-Anything | Skill-driven interactive KG from a codebase. |
Gluecharm / EasySpecs angle
Gluecharm combines a Neo4j view of the codebase with SRS dimensions—overlapping this category with seed manufacturing. EasySpecs brings that story to product teams as the upstream layer feeding agent factories.
Related
References
- Augment — Context engine
- GitLab Knowledge Graph — GitLab Duo
- Bevel — JVM KG
- FalkorDB Code Graph — Graph queries
- GraphGen4Code — IBM research
- Aider repo map — PageRank map
- Understand-Anything — Claude skill
- Gluecharm — SRS + KG