What problem does it solve?
It prevents large AI-coding workspaces from becoming disconnected piles of skills and docs where agents cannot reliably discover, load, or trust the right context.
Core Features & Use Cases
- Defines a four-graph context architecture: skill knowledge graph, document routing graph, memory index, and script/command registry, including the cross-graph edges that enable end-to-end propagation.
- Enforces edge-type discipline with health metrics: models and caps
adjacent, boundary, and verify_with relations to avoid hub-and-spoke degeneration and reduce orphaned/unreachable skills.
- Provides deterministic graph-synthesis signals and propagation discipline: specifies reproducible inputs (frontmatter relations, routing bundle co-membership, explicit labels/keywords) and a checklist to trace a single change across docs, skills, memory, and scripts.
- Covers anti-patterns and remediation prioritization: flags edge inflation, one-way edges, AI-inferred rebuild drift, and ignored cross-graph edges, then remediates orphans in blast-radius order.
Quick Start
Use context-graph when auditing your workspace’s multi-graph context topology to diagnose orphan risk, connectivity failures, and change-propagation gaps.
