Critical failures across naming, descriptions, and schemas. Tool names use wildcard patterns (customers_*, segments_*) rather than explicit verb_noun naming, making them unparseable by LLMs. No input schemas visible in source code for any of the 19 tools. Descriptions are generic placeholders ('Customer management procedures...') under 50 characters, violating the 10-1024 character guideline and providing zero actionable context. No parameter documentation visible. No output schemas documented. The server exposes 19 tools but provides almost no machine-readable or LLM-optimized metadata to guide tool selection or usage.

• Replace wildcard tool names with explicit verb_noun patterns. Instead of 'customers_*', register discrete tools: 'list_customers', 'create_customer', 'get_customer', 'update_customer', 'delete_customer'. Each tool name must start with an action verb and clearly state what happens.
• Add comprehensive input schemas to every tool. Define parameter types (string, number, boolean, enum), constraints (min/max, regex patterns), and descriptions (50-100 chars each). Use JSON Schema with 'type', 'description', 'enum', 'pattern', 'minLength', 'maxLength' fields.
• Rewrite tool descriptions to 50-200 characters, answering: What does it do? When should the LLM call it? What does it return? Example: 'List all customers with pagination. Returns customer ID, name, email, and creation date. Use this to find a customer before updating or deleting.'
• Document output schemas for every tool. Specify return type (object, array), field names, types, and whether fields are required. Include IDs and references needed for downstream tool calls (e.g., customer_id for update_customer).
• Migrate from STDIO to HTTP+SSE or Streamable HTTP transport. STDIO is not remotely testable and cannot integrate with hosted MCP clients. HTTP enables production deployments and hosted agent platforms.