This Go MCP server implements 4 file and shell operation tools with explicit schemas and descriptions. Naming follows verb_noun conventions (execute_shell_command, show_file, search_in_file, write_file). All tools have descriptions of 50-200 chars and structured input parameters with types and per-param descriptions visible in the Go struct tags. However, output schemas are NOT documented in the code provided, only inferred from Go struct field names. Error handling is minimal; security features like path validation exist but are not surfaced in tool descriptions. Tool composition is reasonable (single responsibility per tool), but descriptions lack context about when to use each tool instead of alternatives, and no dry-run/confirmation pattern for destructive operations (execute_shell_command, write_file).

• Add explicit JSON Schema output documentation to each tool. Document the exact fields returned by each tool (e.g., ExecuteShellCommandResult returns {stdout, stderr, exit_code, command, success}). Either add to tool descriptions or create a separate schema registry in code.
• Expand execute_shell_command description from 107 to 150-180 chars: 'Execute a shell command and return stdout, stderr, and exit code. Use this to run system operations, scripts, or command-line tools. Exit code 0 = success; non-zero = error. Stderr is always captured. Command runs in working_dir if provided.'
• Add confirmation pattern for destructive operations. For write_file, add a 'dry_run' boolean parameter (default false). When true, return a preview of what would be written without modifying the file. Guidance: 'Set dry_run=true to preview changes before writing.'
• Document path access restrictions in all file operation tool descriptions. Append: 'File paths must be within configured allowed paths. Denied paths are blocked by the server. Check error messages for available paths if access is denied.'
• Disambiguate show_file vs. search_in_file in descriptions. show_file: 'Retrieve the full contents or a specific line range of a file.' search_in_file: 'Search for a regex pattern in a file and return matching lines with context. Use this when looking for specific content; use show_file when you want to read the entire file or a range.'
• Add error guidance to tool descriptions. E.g., execute_shell_command: 'If the command times out, the error includes how long it waited. If a file is not found, the error lists available files in the directory. If exit_code is non-zero, check stderr for the failure reason.'