Server has 12 tools with complete input schemas and descriptions in Chinese. All tools follow verb_noun naming (list_*, read_*, git_*, analyze_*, review_*). Schemas are well-structured with proper types and required fields. However, descriptions lack depth, most are 20-60 chars, below the 194-char baseline for A+ tools. No output schemas documented. No error handling guidance. No parameter constraints (enums, ranges, patterns). Missing idempotency hints and dependency documentation. Tools are read-only (low risk) but lack recovery guidance for common failures.

• Add output schemas to all 12 tools. Document return types, field names, and structure. E.g., list_repos should return {repos: [{id, name, description, url}], total_count}.
• Expand descriptions to 100-200 chars. Include WHEN to use (e.g., 'Call after list_repos() to see available branches'), WHAT it returns, and any prerequisites. Use English or consistent language.
• Add parameter constraints: enums for 'ref' (branch/tag/commit), min/max for 'max_count' (1-1000), regex pattern for 'pattern' in git_grep. Document in descriptions.
• Implement pagination: add 'limit' (default 50, max 100) and 'offset' or 'cursor' to list_repos, list_branches, git_log, git_grep. Return total_count and next_cursor in responses.
• Add error recovery hints to descriptions. E.g., 'If repo_id not found, call list_repos() to discover available repositories.' Categorize errors as retryable (404 repo) vs fatal (auth failure).

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

• Relies on Roots (deprecated) - pass files/dirs via tool parameters, resource URIs, or configuration