Four tools with complete schemas and descriptions. Tool names follow verb_noun pattern (ado_search_code, ado_get_file_content, ado_get_search_status, ado_detect_context). All have descriptions (avg 150 chars) and input schemas with Zod validation. However, output schemas are undocumented, responses return JSON text blobs without declared field structure. Error handling returns generic error messages without recovery guidance. Parameters lack constraints (e.g., 'top' accepts any number despite claiming max 1000). No pagination support despite search potentially returning large result sets.

• Document output schemas formally. Define the structure of search results (count, results array with fileName, path, repository, repositoryId, project, branch, commitId, matchCount, matches array). Use JSON Schema or TypeScript interfaces in comments.
• Add pagination to ado_search_code: include 'total_count' in response and support 'offset' parameter (or 'next_cursor') to enable iteration through large result sets.
• Enhance error messages with recovery guidance. E.g., 'Repository not found. Call ado_detect_context() to identify valid repositories in your workspace.'
• Add Zod constraints: .number().min(1).max(1000) for 'top' parameter. Validate branch, path, and organization names against ADO naming rules.
• Add dependency hints to parameter descriptions: 'repositoryId (GUID from ado_search_code results)' and 'Call ado_detect_context() first if you don't know your organization/project/repository.'