Agent Skills
› open-mercato/open-mercato
› resolution-playbook
resolution-playbook
GitHub根据工单文本及历史数据,结合决策规则选择最佳处理动作(如设置优先级、分配专家或发送宏回复),确保高风险情况得到优先处理并明确决策依据。
Trigger Scenarios
客服工单路由决策
支持工单分类与分级
Install
npx skills add open-mercato/open-mercato --skill resolution-playbook -g -y
SKILL.md
Frontmatter
{
"name": "resolution-playbook",
"description": "How to choose one resolution action from a ticket and its customer history."
}
Use this playbook to pick the single best action for a support ticket, combining the ticket
text with the history returned by the lookup_ticket_history local tool.
Actions you may propose:
set_priority—payload.priorityone oflow | medium | high | urgent.assign_specialist—payload.team(e.g.billing,security,infra,success).send_macro—payload.macroId(a canned reply, e.g.password-reset,refund-status).
Decision rules:
- Outage, data loss, security, or money-at-risk wording →
set_priority: urgent(this beats every other signal). vip: trueORchurnRisk: highwith an unresolved problem →assign_specialistto the most relevant team; do not let a VIP wait in the general queue.- A routine, well-understood request (password reset, "where is my refund", how-to) with
churnRisk: lowand no open tickets →send_macrowith the matching canned reply. - Everything else →
set_priorityat the level the ticket severity and history justify (more open tickets / slower average resolution → raise it).
When the history and ticket disagree, prefer the safer (higher-touch) action and LOWER your confidence rather than guessing. Always name the deciding signal in the rationale.
Version History
- ab23d45 Current 2026-09-22 10:24


