MCP server for file operations with Git command support
Server has 7 tools with basic schemas and descriptions, but significant gaps limit production readiness. All tools have descriptions (10-50 chars, mostly short), and input schemas are present with Zod validation. However, descriptions lack depth, most are single-sentence statements without context on WHEN to use the tool or what happens on failure. Parameters are typed but lack detailed constraints (e.g., no guidance on path format, no enum for shell choices despite being present in schema). Output schemas are not documented, responses are text-only without structured field definitions. Error handling returns plain text but lacks recovery guidance or error classification. The git_command tool is particularly risky: it accepts arbitrary shell commands with minimal validation, no dry-run option, and no permission gates. No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. Composition is reasonable, tools are single-purpose, but set_base_directory and get_base_directory could be combined into a single stateless operation.
Delete a file or directory
Get the current base directory for file operations
Execute a git command in the base directory
List files in a specified directory
Read the content of a file
Set the base directory for file operations
Write content to a file
Descriptions are too brief (10-50 chars) and lack context. Missing WHEN to use each tool, prerequisites, and side effects. E.g., 'Set the base directory for file operations' does not explain that this affects all subsequent file operations or that it requires an absolute path.
No output schemas documented. Responses are plain text strings without structured field definitions. LLMs cannot plan downstream tool calls or extract typed data. E.g., list_files returns a text blob instead of {files: [{name, isDirectory}], total}.
git_command tool accepts arbitrary shell commands with minimal validation. No dry-run option, no confirmation step for destructive operations, and no permission gates. Agents can execute 'git reset --hard' or 'git push --force' without safeguards.
Inferred effective spec: <=2025-11-25.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 47 | <=2025-11-25 | v2 |
No tool annotations (readOnlyHint, destructiveHint, idempotentHint) despite clear risk levels. delete_file and write_file are destructive; git_command is high-risk. Agents cannot reason about side effects or retry safety.
Error handling returns plain text without recovery guidance or error classification. E.g., 'Error setting base directory: EACCES' tells the agent nothing about whether to retry, ask the user, or give up. Missing pattern:recovery-guide.