Strong foundation with 27 well-named tools, explicit schemas, and tool annotations. All tools have descriptions and input schemas with typed parameters. However, descriptions are often terse (many under 50 chars), output schemas are undocumented, and error handling guidance is absent. Parameter descriptions lack detail on constraints, formats, and dependencies. No pagination guidance for list tools. Composition is sound, tools are single-responsibility and chainable via IDs.

• Document output schemas for all tools. Add a 'returns' field to each JettyToolDefinition describing the structure (e.g., list-collections returns {collections: [{name, description, env_keys}], total_count}). This enables LLM planning and chaining.
• Expand descriptions to 100-200 chars. Add WHEN/WHY context. E.g., 'List all collections' → 'List all collections in your workspace. Call this first to discover available collections before creating tasks or running workflows.'
• Add constraint details to parameter descriptions. E.g., limit: 'Max results per page (1-100, default 10)'; cadence: 'Cadence config. Examples: {type:"hourly", minute_utc:15}, {type:"daily", hour_utc:9, minute_utc:0}. Required fields depend on type.'
• Add error recovery hints to descriptions. E.g., get-task: 'Get task details. If task not found, call list-tasks(collection) to see available tasks.' This guides multi-step agent planning.
• Document pagination behavior. E.g., list-trajectories: 'Returns up to `limit` results (default 10, max 100). Use `page` to fetch subsequent pages. Response includes total_count for UI pagination.' Specify max result caps to prevent context window exhaustion.