Eight tools with complete JSON schemas and descriptions. Naming follows verb_noun pattern (read_file, list_dir, grep, read_blocks, edit_block, create_file, git_diff, exec). Descriptions are clear and action-oriented (avg ~100 chars), explaining WHAT each tool does and WHEN to use it. All parameters have type definitions and descriptions. However, parameter descriptions lack constraint details (ranges, formats, allowed values). Error handling guidance is minimal, no recovery hints or actionable error messages documented. Output schemas are not explicitly documented. Missing tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk profiles (READ_ONLY vs WRITE). No pagination or result-limiting guidance for grep/list_dir.

• Add tool annotations to all tools: readOnlyHint=true for read_file, list_dir, grep, read_blocks, git_diff; destructiveHint=true for edit_block, create_file, exec. This enables MCP clients to apply safety policies and warn before execution.
• Enhance parameter descriptions with explicit constraints: 'path: string, relative to project root, max 1024 chars, no ../ traversal'; 'limit: integer, 1 - 10000, default 100'; 'pattern: literal substring (no regex), max 500 chars'.
• Document output schemas for each tool. E.g., read_file returns {lines: [{number: int, text: string}], total_lines: int}; grep returns {matches: [{path: string, line_number: int, text: string}], total_matches: int}.
• Add error handling guidance to descriptions. E.g., edit_block: 'Returns error if target_hash is stale (file changed since read_blocks). Call read_blocks again to get current hash.' grep: 'Returns empty matches array if pattern not found, this is not an error.'
• Add pagination guidance to grep and list_dir. E.g., 'Results capped at 100 matches. For larger searches, narrow the path parameter or use multiple grep calls with different patterns.'