Codebase Knowledge Graph (codebase-memory-mcp)

This project uses codebase-memory-mcp to maintain a knowledge graph of the codebase.

Routing Boundary

Select this skill by default when the material question requires exact structural evidence from code: symbol discovery, static callers/callees, cross-file dependencies, code architecture, change impact, or cross-service paths. Once selected, prefer its MCP graph tools over broad grep/glob/file scanning.

Use $repowise for behavioral location, repository health, risk, history, or dead-code triage; $codegraph for implicit runtime dispatch; $graphify for broad corpus/community exploration; and rg or direct reads for literals, configuration, Markdown/policy, or graph coverage gaps.

Priority Order

1. search_graph — find functions, classes, routes, variables by pattern
2. trace_path — trace who calls a function or what it calls
3. get_code_snippet — read specific function/class source code
4. query_graph — run Cypher queries for complex patterns
5. get_architecture — high-level project summary

• detect_changes — Map git diff to affected symbols + blast radius with risk classification.
• get_graph_schema — Node/edge counts, relationship patterns, property definitions per label.
• search_code — Grep-like text search within indexed project files.
• manage_adr — CRUD for Architecture Decision Records.

When to fall back to grep/glob

• Searching for string literals, error messages, config values
• Searching non-code files (Dockerfiles, shell scripts, configs)
• When MCP tools return insufficient results

Examples

• Find a handler: search_graph(name_pattern=".*OrderHandler.*")
• Who calls it: trace_path(function_name="OrderHandler", direction="inbound")
• Read source: get_code_snippet(qualified_name="pkg/orders.OrderHandler")