Critical failures across all dimensions. Source code inspection reveals tool definitions exist in server.py but NO VISIBLE INPUT SCHEMAS, parameter descriptions, or output documentation. All 41 tools have only bare names and risk labels. Descriptions are trivial (10-20 chars: 'Acknowledge a pinned message', 'Add a role to a channel'). No parameter type definitions, constraints, or LLM guidance. No error handling documentation. This server exposes raw domain logic without the abstraction layer agents need. STDIO transport compounds the issue, not remotely testable.

• Add full JSON Schema input definitions to every tool. Include type, description, enum, minLength, maxLength, pattern, and default for each parameter. Example: { 'channel_id': { 'type': 'string', 'description': 'The channel ID (e.g., 'general'). Use list_channels() to discover available channels.', 'minLength': 1 } }
• Expand tool descriptions to 50-200 characters. State WHAT the tool does, WHEN to use it, and what it returns. Example: 'Send a message to the channel. Returns message_id and timestamp. Use this to post updates, questions, or replies. Requires channel_id and message content.'
• Document output schemas for all tools, especially list_* and get_* variants. Specify returned fields, types, and which IDs are safe to pass to downstream tools. Example: list_messages returns [{ message_id, channel_id, sender, content, timestamp }].
• Add error handling guidance to tool descriptions. Example: 'If channel not found, call list_channels() to see available options. If permission denied, verify your role via list_roles().'
• Implement confirmation/dry-run for destructive tools (delete_channel, delete_message). Add a 'confirm' boolean parameter (default false) that returns a preview before executing. Pair with undo_* tools for recovery.