Antigravity exposes 3 filesystem tools with minimal documentation. Input schemas are visible and typed, but lack depth. No output schemas documented. No error handling guidance. No parameter descriptions beyond the bare type. The 'filesystem.write' tool is destructive but has no confirmation pattern, dry-run option, or warning in its description. No tool annotations (readOnlyHint/destructiveHint). Overall, definitions are skeletal and would not guide an LLM reliably.

• Expand tool descriptions to 100-200 chars. Example: 'filesystem.list: List files and directories in a given path. Returns an array of file/directory names with metadata (size, type, modified date). Use this to explore directory structure before reading or writing files.'
• Document output schemas for all tools. filesystem.list should return {type: 'array', items: {type: 'object', properties: {name, type, size, modified}}}. filesystem.read should return {type: 'object', properties: {content, encoding, size}}.
• Add parameter descriptions. 'path': 'Absolute or relative filesystem path. Must not contain .. or symlinks outside the allowed root. Max 1024 chars.' 'content': 'File content as UTF-8 string. Max 10MB.'
• Add destructiveHint annotation to filesystem.write and implement a dry-run parameter or confirmation step before overwriting existing files.
• Add error handling guidance: 'If path not found, check spelling and permissions. If permission denied, verify agent has read/write access to the directory.'