Hivemind exposes 4 tools for inter-agent communication. All tools have descriptions and basic input schemas, but schemas lack parameter type definitions and descriptions. Tool names are action-oriented (ask_, wait_for_, get_) but descriptions are terse (avg 95 chars, below 194-char baseline). No output schemas documented. Parameters like 'text' and 'timeout_ms' lack descriptions in the schema itself. Error handling is minimal, no recovery guidance or categorization. The tools are READ_ONLY (low risk) but lack the structure expected of production tools.

• Add type and description to every parameter in the input schema. E.g., ask_codex schema should include: 'text': {type: 'string', description: 'Prompt to send to Codex. Keep under 5000 chars.'}, 'wait_ms': {type: 'number', description: 'Milliseconds to block (0=fire-and-forget, 60000-120000 recommended). Min: 0, Max: 300000.'}
• Document output schemas for each tool. E.g., wait_for_codex returns: {type: 'object', properties: {messages: {type: 'array', items: {type: 'object', properties: {id, timestamp, text, sender}}}}}
• Expand tool descriptions to 100-200 chars with WHEN/WHY guidance. E.g., peers: 'List peers visible to Hivemind (currently Codex). Call this first to verify Codex is online before sending prompts. Returns threadId, online status, and warnings (e.g., unsupported_session_picker).'
• Add error handling with recovery hints. E.g., if ask_codex times out: 'Codex did not reply within 120s. Codex may be offline or overloaded. Try wait_for_codex() to check for buffered messages, or restart Codex with `hm codex`.'
• Constrain numeric parameters: timeout_ms should have min: 0, max: 300000; wait_ms should have min: 0, max: 300000. Document these in descriptions.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry