Token导航 LogoToken导航

Agent Skill · 研究检索

support(support 搜索)

support 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

support 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为权限未标注;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 16,847。

总安装 6,957周安装 287GitHub Stars 公开资料未说明下载 2,273核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
1.0.0

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:support(support 搜索)
来源仓库:https://github.com/agenticio/support
安装命令:
openclaw skills install support
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install support

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于起草对客户问题的答复,确保语气和技术准确性。适合客户支持场景的快速响应。可整合知识库和常见问题解答。需确认问题背景和用户身份。support 属于研究检索类 Skill,可作为该场景下的辅助能力补充。建议审核回复内容后再发送。适用宿主包括 OpenClaw,接入前应确认版本、权限和运行环境要求。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install support;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    权限没有明确标注,第一次使用时只给一个无关紧要的任务,观察它到底读了哪些文件和网络。

常见问题

6 条

support 怎么安装?

在终端执行 openclaw skills install support。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

support 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

support 需要什么权限?

权限级别没有明确标注。第一次使用时建议只给无关紧要的任务,观察它实际读写了哪些内容。

support 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

support 用的人多吗?

公开数据是累计安装 6,957 次,近一周安装 287 次,下载 2,273 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

support 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
support

support 搜索 · 本页

6,957—权限未标注MIT-0通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

name
Support
description
A comprehensive AI agent skill for delivering exceptional customer support. Drafts responses to customer issues with the right tone and technical accuracy, manages support queues by priority, identifies patterns in recurring problems, escalates complex cases appropriately, maintains consistency across every interaction, and helps support teams and solo founders provide enterprise-quality customer experience without enterprise headcount.

Support

The Moment That Defines Your Relationship

A customer has been trying to solve a problem for forty minutes. They have read the documentation. They have tried three things that did not work. They have searched your help center and found an article that seemed relevant and turned out not to be. They are now writing to you, and they are not in a neutral state. They are frustrated. They feel like they have wasted time. They are questioning whether they made the right decision using your product.

What happens in the next few hours will determine whether this person becomes a loyal customer who tells others about you, a neutral customer who continues using your product without enthusiasm, or a lost customer who cancels and leaves a review that reflects how they felt at their most frustrated moment.

The response they receive will make that determination. Not your product features. Not your pricing. Not your marketing. The response.

Support exists to make sure that response is exactly right every time — not just when your best support person is fully rested and fully focused, but at 11pm on a Friday when the queue has backed up and the person responding has already handled thirty tickets today.


Understanding What Customers Actually Need

The surface request in a support ticket is rarely the whole picture. A customer who writes "this feature is broken" may be experiencing a genuine bug, a misconfiguration, a misunderstanding of how the feature works, or a gap between what the feature does and what they expected it to do. Each of these requires a completely different response. Treating them all as bug reports wastes everyone's time and erodes the customer's confidence.

The skill reads the full ticket — not just the subject line, but the description, the account history, the context clues about the customer's technical level and emotional state — and identifies what is actually going on before drafting any response.

It distinguishes between the customer who needs a technical solution, the customer who needs an explanation, the customer who needs reassurance that their frustration is valid and that someone is taking their problem seriously, and the customer who needs a combination of all three. It calibrates the response accordingly.

This matters because customers are not evaluating your support on technical accuracy alone. They are evaluating whether they felt heard. A technically correct response delivered in a tone that feels dismissive fails the interaction even if it solves the problem. A warm, empathetic response that does not actually solve the problem fails in a different way. The skill optimizes for both simultaneously.


Response Quality at Scale

Individual support interactions are easy to do well when volume is low and the person responding has time and energy. The challenge is maintaining that quality when volume is high, when the same questions come in repeatedly, when the person responding is fatigued, and when the pressure to clear the queue creates shortcuts that degrade quality.

The skill solves this by separating the cognitive work from the communication work. The cognitive work — understanding what the customer needs, identifying the solution, determining the right escalation path — requires judgment and cannot be fully automated. The communication work — translating that understanding into a clear, warm, professionally written response — can be done consistently regardless of queue volume or time of day.

When you identify what a customer needs and what the resolution is, the skill produces the response. Not a template with fill-in-the-blank fields. An actual response that addresses the specific customer's specific situation in language that feels personal and considered rather than copied from a library.

For common issues that you handle repeatedly, it helps you build a knowledge base of solutions that improves the quality and speed of future responses — not by sending canned replies, but by ensuring that the solution knowledge is always available and correctly applied.


Tone and Voice Consistency

Every customer support interaction is a brand touchpoint. The tone of your support — whether it is warm or cold, formal or casual, patient or hurried — communicates something about your company that marketing cannot override.

Most support teams struggle with tone consistency. Individual responders have individual styles. Different people handle different ticket types with different registers. Responses at the beginning of the day are calmer than responses at the end. Responses to easy questions are warmer than responses to difficult ones.

The skill maintains consistent tone across every interaction. You define what your support voice sounds like — the level of warmth, the degree of formality, the approach to technical language, the way you handle frustration — and every response reflects that voice regardless of who is handling the ticket or when.

For solo founders handling support personally, this means your support voice remains consistent even when you are tired, stressed, or handling a particularly difficult customer. For teams, it means every team member sounds like the same company even if their individual communication styles are very different.


Priority and Queue Management

Not all support tickets are equally urgent. A customer who cannot access their account at all is in a different situation from a customer who has a question about a feature they have not tried yet. A customer whose paid workflow is completely blocked is in a different situation from a customer who encountered a minor inconvenience.

The skill triages incoming tickets by urgency based on the impact to the customer, the nature of the issue, the customer's account status and relationship history, and any signals about emotional state that suggest an elevated response is warranted. It surfaces the highest-priority items regardless of when they arrived in the queue.

It also identifies when a ticket that looks simple is actually complex — when the described problem is a symptom of something larger, when the customer's account history suggests this is a recurring issue rather than a one-time occurrence, when the resolution requires input from engineering or product rather than just support.


Pattern Recognition

Individual support tickets are problems to solve. Patterns across many support tickets are product problems to fix.

When the same question comes in repeatedly, it means either the product is confusing in a specific way or the documentation is failing to answer a question customers frequently have. When the same error appears across multiple accounts, it may indicate a bug or a configuration issue that affects a class of users. When a specific feature generates disproportionate support volume, it signals either a usability problem or a gap between customer expectations and what the feature actually does.

The skill tracks these patterns across your ticket history and surfaces them in regular summaries. Not raw data. Actionable insights: the three questions that account for thirty percent of your support volume, the feature that generates the most confusion, the onboarding step where customers most frequently get stuck.

These insights are the feedback loop between support and product. They turn support from a cost center that reacts to problems into an intelligence function that prevents them.


Escalation

Some support issues cannot be resolved by support alone. Bugs that require engineering. Billing disputes that require account management. Security concerns that require immediate escalation outside normal channels. Feature requests that need to reach product. Legal concerns that need to reach the appropriate team immediately.

The skill identifies when a ticket requires escalation, determines the appropriate escalation path, and produces the internal summary that gives the receiving team everything they need to continue without requiring the customer to repeat themselves. The customer experience of a good escalation should be seamless: they feel like everyone they speak to is fully informed about their situation, not like they are starting over every time they reach a new person.


For Solo Founders

For founders who handle support personally — which is most founders at some point — support is the most direct window into how customers actually experience your product. It is also the most time-consuming and most emotionally demanding part of the job, particularly when the customer is frustrated and the problem is complicated.

The skill helps you handle support sustainably. It reduces the time per ticket without reducing the quality. It helps you maintain warmth and patience in your responses even when you are tired or the queue is long. It builds the knowledge base that makes future tickets faster. And it surfaces the product insights buried in your support history that will reduce the volume of tickets over time.

Support done well is not a burden. It is the most reliable source of product intelligence you have, and the most direct way to build the customer loyalty that drives word-of-mouth growth. The skill makes it possible to do it well consistently, not just when conditions are ideal.

平台分布

OpenClaw

70.28%
按下载量换算1,597

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

权限未标注

当前来源未能明确判断权限范围,默认进入异常复核队列。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills