一键导入
after-sales-qa
基于知识库帮助用户解决售后相关问题,包括退换货、退款、物流、质量问题等 Use for qa tasks in the general domain for external_agent users, with platform RAG grounding and citations.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
基于知识库帮助用户解决售后相关问题,包括退换货、退款、物流、质量问题等 Use for qa tasks in the general domain for external_agent users, with platform RAG grounding and citations.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
根据用户提供的简历文本和岗位JD,评估候选人与岗位的匹配度,分析亮点与不足,并给出建议 Use for case_triage tasks in the hr domain for manager users, with platform RAG grounding and citations.
当用户询问机械产品制造工艺、加工步骤、工艺要求或工艺知识时使用,基于机械产品工艺知识库回答并返回引用。 Use when an agent needs to invoke the published HTTP API Skill, follow its input/output schema, or return knowledge-grounded answers with citations.
基于机械产品工艺知识库,回答关于材料热处理、焊接工艺评定、切削参数等工艺问题,提供具体参数、步骤说明并引用知识库原文 Use when an agent needs to invoke the published HTTP API Skill, follow its input/output schema, or return knowledge-grounded answers with citations.
基于机械产品工艺知识库,查询工艺参数、规范标准、材料选择建议和加工流程步骤。当知识库中无相关信息时返回空结果。 Use when an agent needs to invoke the published HTTP API Skill, follow its input/output schema, or return knowledge-grounded answers with citations.
| name | after-sales-qa |
| description | 基于知识库帮助用户解决售后相关问题,包括退换货、退款、物流、质量问题等 Use for qa tasks in the general domain for external_agent users, with platform RAG grounding and citations. |
Use this skill when the user needs help with this enterprise task:
面向消费者的售后问题解答 Skill。用户遇到退换货、退款、物流异常、商品质量等售后问题时,通过本 Skill 获取基于知识库的准确解答。输出需引用知识库原文,若知识库无匹配内容则引导用户转人工客服。
This skill is backed by the knowledge-base platform runtime. The knowledge base is a resource dependency; the task scenario is the reason to invoke the skill.
references/task-scenario.md if you need more detail on the domain, intent, audience, or output style.references/api.md before calling the runtime endpoint.SKILL_API_KEY.http://localhost:3000/api/public/skills/after-sales-qa/runPOSTReturn the API answer directly. If the response includes citations, surface the most relevant sources. If the API reports no reliable knowledge-base evidence, say that clearly instead of filling gaps.