Server has 9 tools with basic schemas and descriptions, but significant gaps limit production readiness. All tools have descriptions (10-80 chars, mostly adequate) and input schemas with types. However, output schemas are completely undocumented, LLMs cannot plan downstream calls or extract results. Parameter descriptions are minimal (e.g., 'The project name in the graph' repeated 6 times). No error handling guidance, no pagination support despite search/query tools returning potentially large result sets, and no tool annotations (readOnlyHint/destructiveHint). The index_repository tool is marked WRITE but has no confirmation pattern or dry-run option. Tool naming is verb-forward (get_, search_, trace_, query_, list_, index_) which is good, but descriptions lack WHEN/WHY context and actionable guidance.

• Document output schemas for all tools. Example for search_graph: 'Returns {results: [{id: string, label: string, name: string, properties: object}], total: integer, limit: integer}'.
• Add pagination to search_graph and query_graph: include offset/limit in request, return total_count and has_more in response. Cap results at 50 items by default.
• Enhance parameter descriptions with format hints and discovery guidance. E.g., 'project: string, project name (call list_projects to see available projects; case-sensitive)'.
• Add tool annotations: mark all read-only tools with readOnlyHint=true, mark index_repository with destructiveHint=true, mark idempotent tools accordingly.
• Implement confirmation pattern for index_repository: add optional 'dry_run' parameter and return a summary of changes before committing. Or require explicit 'confirm=true' parameter.

Inferred effective spec: <=2025-11-25.

• Relies on Stateful initialize / Mcp-Session-Id (removed; protocol is stateless) - make each request self-contained