Server defines 6 tools with complete JSON schemas and descriptions. Naming follows verb_noun convention (list_, ask_, get_, health). Descriptions are present (avg ~120 chars) but lack actionable context for LLM selection. Parameters are typed but many lack descriptions (context array in ask_agent has no item-level guidance). No output schemas documented. Error handling and recovery guidance absent. Security considerations (agent isolation, permission checks) not visible in tool definitions.

• Document output schemas for all tools. E.g., list_internal_agents should specify: returns {agents: [{id: string, name: string, description: string}], total: number}.
• Add descriptions to ask_agent's context array items: 'Array of prior conversation turns. Each item: {role: "user"|"assistant", content: string (message text)}'.
• Add error recovery guidance to ask_agent and ask_external_agent: 'If agent is offline or ID invalid, call list_internal_agents or list_agents to verify availability.'
• Add limit and offset parameters to list_internal_agents and list_agents with defaults (e.g., limit=20, offset=0) and document max result count.
• Clarify when to use ask_agent vs ask_external_agent in their descriptions: 'ask_agent: internal Memoh bots. ask_external_agent: curated external A2A agents. Use list_* first to discover.'

Inferred effective spec: 2026-07-28+.

• Current-spec patterns present: Streamable HTTP