OpenClaw Codex

模板库 / Templates

3 个重点场景模板(中文 / English)

模板只给结构,不给真实值。所有字段都使用占位符,可用于公开内容、团队协作和快速起步。

Templates are structure-first and secret-free. Every field uses placeholders so you can publish safely and iterate quickly.

模板目录 / Template Index

与场景页一一对应:客服、社群、告警值班。

模板 1:客服与工单 / Support & Ticket Playbook

中文:适用于客服咨询、售后问题和重复工单处理。

EN: For support inquiries, after-sales issues, and repetitive ticket handling.

name: support_ticket_playbook
input:
  channel_event: <CHANNEL_EVENT>
pipeline:
  - normalize_message
  - classify_intent
  - check_faq_index: <FAQ_INDEX>
  - decide_handoff:
      conditions: [high_risk, repeated_issue, low_confidence]
  - create_or_update_ticket: <TICKET_SYSTEM>
output:
  auto_reply: <RESPONSE_TEMPLATE>
  ticket_id: <TICKET_ID>
metrics:
  - fcr
  - frt
  - handoff_rate
  - escalation_rate

模板 2:社群运营 / Community Operations Playbook

中文:适用于社群欢迎流、规则执行、FAQ 与风险词分流。

EN: For community onboarding, policy enforcement, FAQ handling, and risky-content routing.

name: community_ops_playbook
input:
  group_event: <GROUP_EVENT>
pipeline:
  - welcome_new_member
  - run_policy_checks:
      ruleset: <RULESET_ID>
  - faq_match:
      index: <FAQ_INDEX>
      fallback: human_mention
  - moderation_route:
      actions: [warn, mute, escalate]
output:
  community_action: <ACTION_RESULT>
metrics:
  - d7_retention
  - faq_hit_rate
  - manual_intervention_count
  - active_member_ratio

模板 3:告警值班 / Incident Response Playbook

中文:适用于监控告警接入、降噪路由和值班协同。

EN: For monitoring alert ingestion, noise reduction routing, and on-call collaboration.

name: incident_response_playbook
input:
  alert_event: <MONITORING_ALERT>
pipeline:
  - dedupe_alert:
      key: [service, fingerprint, severity]
      window_minutes: <DEDUP_WINDOW_MINUTES>
  - route_by_severity:
      p1: page_primary_oncall
      p2: notify_incident_channel
  - ack_and_close_loop:
      ack_writeback: true
      resolve_writeback: true
  - draft_postmortem_stub
output:
  incident_status: <INCIDENT_STATE>
metrics:
  - mtta
  - mttr
  - alerts_per_hour
  - false_positive_rate

脱敏检查 / Sanitization Checklist

  • 不要公开 token、API Key、签名密钥、Webhook 密钥。
  • Do not publish tokens, API keys, signing secrets, or webhook secrets.
  • 不要公开服务器 IP、内部域名、端口映射、真实目录。
  • Do not expose server IPs, internal domains, port mappings, or real paths.

配套入口 / Related Pages