11 tools with consistent naming (verb_noun pattern: list_, get_, chutes_*). All have descriptions (avg ~80 chars, within baseline 34-392). Input schemas present for all tools with proper types and defaults. However, descriptions are terse and lack LLM-optimized guidance on WHEN to use each tool or what to do with results. Parameter descriptions are minimal (e.g., 'Page number for pagination' lacks range constraints). Output schemas are not documented, LLMs cannot plan downstream calls. Error handling is absent from tool definitions. No tool annotations (readOnlyHint/destructiveHint). chutes_deploy_vllm is marked [BETA] but lacks confirmation/dry-run pattern for a destructive write operation.

• Document output schemas for all tools. E.g., chutes_list_models should return {models: [{id, name, owned_by, created_at}], total_count, has_more}. This enables LLMs to plan chaining.
• Add min/max constraints to numeric parameters. E.g., 'limit: 1 - 100 (default 25)', 'page: 0 - N (zero-indexed)', 'max_tokens: 1 - 4096 (default 512)'.
• Expand tool descriptions to 80 - 150 chars with WHEN guidance. E.g., 'List stable model aliases configured for this account. Call this to see which model names map to specific versions before invoking chutes_chat_complete.'
• Add tool annotations: mark all read-only tools with readOnlyHint=true; mark chutes_deploy_vllm with destructiveHint=true and idempotentHint=false. Implement a confirmation pattern (e.g., 'dry_run' param or separate 'confirm_deploy' tool).
• Add error recovery guidance to tool descriptions. E.g., 'If authentication fails, verify CHUTES_API_KEY is set. If quota exceeded, call chutes_get_quota() to check limits.'