Server has 5 tools with basic schemas and descriptions, but significant gaps limit production readiness. All tools have descriptions (10-50 chars, below the 50-200 char baseline for LLM optimization), and input schemas are present but minimal. Parameters lack detailed descriptions explaining constraints, formats, or dependencies. No output schemas documented. Error handling is absent, tools return generic success messages without recovery guidance. Tool names follow verb_noun convention (list_, start_, complete_, get_) but descriptions are too terse to guide LLM selection. No parameter validation rules, enums, or type constraints visible beyond basic JSON Schema types.

• Expand tool descriptions to 50-200 chars. Example: 'list_workflows' → 'List all available workflow templates. Call this first to discover workflow IDs before starting a session. Returns workflow name, description, and step count.'
• Add parameter descriptions with format/constraint details. Example: 'workflow_id: Unique identifier of the workflow template (alphanumeric, 1-50 chars). Use list_workflows to discover valid IDs.'
• Document output schemas for all tools. Example: 'list_workflows returns {workflows: [{id, name, description, step_count}], total_count}'. This enables LLMs to chain tools correctly.
• Implement structured error responses with recovery guidance. Example: 'Error: workflow_id "invalid-123" not found. Valid workflows: onboarding, approval-process, data-entry. Call list_workflows() to see all available.'
• Add parameter validation and return clear error messages. Validate workflow_id/session_id format early; return 'Invalid workflow_id: must be alphanumeric, 1-50 chars' instead of generic failures.

Inferred effective spec: <=2025-11-25.

• Relies on Logging (deprecated) - log to stderr or use OpenTelemetry