Customer Service

Problem

In customer support, common answers are often scattered across Markdown, JSON, YAML, and TXT files, while responses need to stay accurate and natural. customer-service organizes knowledge-base retrieval, issue classification, templated replies, and ticket recording into a clear workflow for teams that maintain product-line FAQs.

How It Works

• Reads references/config.toml and uses scripts/config_manager.py to initialize or update settings.
• Searches a product-organized knowledge base with scripts/knowledge_search.py, supporting Markdown, JSON, YAML, and TXT.
• Classifies questions into feature inquiries, usage issues, after-sales, technical support, or complaints.
• Generates replies from references/reply_templates.json, or creates a ticket with scripts/create_ticket.py when coverage is missing.
• Ticket fields include ticket_id, product, category, priority, status, question, customer_info, and resolution; priority can be urgent/high/normal/low.

Boundaries

It fits scenarios with structured knowledge bases, standardized support replies, and ticket traceability. If the knowledge base does not cover a question, the skill records a ticket rather than solving it directly; reply tone depends on reply_templates.json and is not suited for strict legal wording, multilingual compliance, or heavily manual review workflows.