Ayiru MCP provides 14 tools with complete JSON Schema input definitions and descriptions. Naming follows verb_noun conventions (ask, validate_command, get_tool_spec, search_tools, etc.). However, descriptions are inconsistent in depth and actionability. Most are 30-80 chars, below the 50-200 char baseline for LLM optimization. Parameter descriptions exist but are minimal (e.g., 'The question to ask' for ask.question). Output schemas are not documented in the visible code. Error handling guidance is absent, tools return HTTP semantics without recovery hints. The submit_claim tool (WRITE) lacks confirmation/dry-run pattern. Tool composition is sound (single responsibility, chainable IDs), but descriptions need expansion to guide LLM selection and parameter usage.

• Expand tool descriptions to 80-150 chars. For each tool, answer: What does it do? When should the LLM call it instead of similar tools? What does it return? Example: 'ask' → 'Retrieve tool documentation and examples via semantic search. Use this to answer questions about tool capabilities, syntax, and best practices. Returns ranked results with snippets and tool_id for chaining.'
• Document output schemas for all tools. Add a 'Returns' section to each tool description listing key fields (e.g., 'Returns: {results: [{tool_id, name, description, risk_level}], total_count, next_cursor}'). This enables LLMs to plan downstream calls.
• Add min/max constraints to numeric parameters and enums to string parameters. E.g., limit: 'integer, 1-100 (default 20)'; claim_type: 'enum: [capability, constraint, effect, risk]'. Constraints prevent invalid LLM inputs.
• Implement dry-run for submit_claim. Add a dry_run: boolean parameter (default false) that validates the claim without persisting it. Return the validated claim with verification status so the agent can preview before committing.
• Add error recovery hints. When a tool fails (e.g., resolve_subject returns 404), include a suggestion: 'Subject not found. Try search_tools() with a partial name, or ask() for general tool discovery.' This guides the LLM to self-correct.