OpenMoat defines 7 tools with complete input schemas and descriptions. Naming follows verb_noun convention (tool_search, tool_describe, tool_call, tool_result_fetch, moat_health, moat_usage, registry_list). Descriptions are present and substantive (50-150 chars), explaining WHAT each tool does and WHEN to use it. However, parameter descriptions lack depth, many omit format/constraint guidance. Output schemas are not documented. Error handling is minimal; no recovery guidance or actionable error messages visible in source. Tool composition is sound (each does one thing), but descriptions could better explain dependencies and prerequisites.

• Document output schemas for all 7 tools. For tool_search, specify: returns array of {server, tool, score, snippet, file, heading}. For moat_health, specify: returns {pool_size, breaker_state, latency_ms, failure_count}. This enables LLMs to chain tools correctly.
• Enhance parameter descriptions with format and constraint guidance. E.g., 'query (1-500 chars): search terms separated by spaces; matches tool names first, then server names, then descriptions' and 'hash (48 hex chars): SHA-384 digest of result archive'.
• Add recovery hints to tool descriptions. E.g., tool_describe: 'If server not found, call registry_list() first to see available servers.' tool_call: 'If tool not found, call tool_describe(server) to list available tools.'
• Clarify moat_health and moat_usage outputs. Specify what 'bounded' means, what fields are returned, and what values indicate health vs. degradation. E.g., 'Returns {pool_size, active_connections, breaker_state: open|half_open|closed, p99_latency_ms, error_rate_percent}'.
• Add idempotency hints. tool_call should state: 'Idempotent if tool is read-only (risk=READ_ONLY). Non-idempotent if tool is write (risk=WRITE), retries may cause duplicate side effects.' This guides agent retry logic.