Two tools with minimal documentation. read_file has a basic description (54 chars) and a single string parameter with minimal context. delete_file has identical description length but no visible input schema in the provided code excerpt. Both tools lack comprehensive parameter descriptions, output schema documentation, and error handling guidance. The server appears to be a test harness fixture, not a production MCP server, which explains the sparse definitions. No tool annotations (readOnlyHint/destructiveHint) despite clear risk profiles.

• Expand read_file description to ~100 - 150 chars: 'Read a UTF-8 text file from the fixture root. Returns the file path and content. Fails if file exceeds 64 KB or is not valid UTF-8. Use this to inspect test fixtures before running security tests.'
• Expand delete_file description and add destructiveHint annotation: 'Permanently delete a regular file from the fixture root. This operation cannot be undone. Use with caution in test cleanup. Returns the deleted file path on success.'
• Add output schema documentation for both tools. read_file should document {path: string, content: string}. delete_file should document {path: string, deleted: boolean}.
• Add parameter constraint descriptions: 'path: Relative path (no ../, no absolute paths, no symlinks). Max 255 chars. Example: fixtures/test.txt'
• Implement tool annotations: read_file with readOnlyHint=true, delete_file with destructiveHint=true and idempotentHint=false (since repeated deletes fail).