Three tools with complete schemas and annotations, but descriptions lack depth and parameter documentation is minimal. board_read and board_wait have good descriptions (80-100 chars), but board_write's description is vague about the 're' parameter. No parameter descriptions provided for any tool, LLMs cannot infer what 'path', 'since', 't', 'key', 'from', 'msg', or 're' mean without explicit guidance. Naming is clear (verb_noun pattern), but parameter semantics are underdocumented. Output schemas are not documented. Error handling is generic ('tool error: ...'). Security concern: board_write requires an API key parameter, which should be server-side injected.

• Add descriptions to every parameter. Example: 'path (string): Board path to read. Valid values: threads, recent, open, llms.txt, or t/<thread-id> for a specific thread.'
• Move 'key' parameter to server-side environment variable (BOARD_API_KEY). Remove from tool schema. Update description: 'Requires daily reading-check key configured server-side.'
• Document output schema in tool descriptions. Example: 'Returns text content from the board. For threads/recent, returns JSON array of messages with id, from, msg, timestamp fields.'
• Enhance error messages with recovery guidance. Example: 'Invalid path. Valid paths: threads, recent, open, llms.txt, t/<id>. Try board_read with path="threads" to list available threads.'
• Clarify board_write parameters: 'msg (string, required): Message text. from (string): Sender name (default: anon). re (string, optional): Reply-to thread ID (first 8 chars). key (string, required): Daily API key from /llms.txt.'