This is a Rust CLI tool, not an MCP server. The 4 tools (get_node, neighbors, resolve, resolve_strict) are graph traversal functions with READ_ONLY semantics. Tool definitions are inferred from source code comments and function signatures, not from explicit MCP tool registration. Input schemas are partially visible but lack proper JSON Schema structure with type constraints. Descriptions exist but are terse (10-50 chars) and lack LLM-optimized guidance on when/why to use each tool. No output schemas documented. No error handling patterns visible.

• Implement this as a proper MCP server using the MCP Rust SDK (or similar). Register tools via the tools/list endpoint with explicit JSON Schema input definitions and documented output schemas.
• Expand tool descriptions to 50-200 chars, explaining WHAT each tool does, WHEN to use it (vs similar tools), and WHAT it returns. E.g., 'resolve: Finds a node by slug, exact label, or substring. Use when you have a user-provided name. Returns node ID or ambiguity error listing candidates.'
• Define input schemas with proper JSON Schema: specify types (string, number, object), constraints (enums, min/max), and descriptions for every parameter. Replace 'ns: NamespaceConfig' with explicit field definitions.
• Document output schemas for each tool. Specify the structure of returned nodes, edges, and error responses. Include pagination parameters (limit, offset, total_count) for tools returning lists.
• Add error handling guidance: 'If resolve returns ambiguity error, call resolve_strict for stricter matching' or 'If get_node returns not-found, try neighbors with a broader search.' This guides LLM recovery.