Token导航 LogoToken导航

Agent Skill · 研究检索

discord-agent-communication(Discord Agent communication 搜索)

用于处理 Discord 服务器、频道、消息、成员和机器人交互。它适合让 Agent 辅助查询社区对话、整理频道内容、发布通知或管理基础协作流程。使用时需要确认 bot 权限、频道可见范围和服务器规则;涉及删除消息、管理成员、批量通知或读取私密频道时,应先获得明确授权并控制操作范围。

discord-agent-communication 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为需要联网;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 7,516。

总安装 18,769周安装 798GitHub Stars 1下载 6,576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
1.0.1

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:discord-agent-communication(Discord Agent communication 搜索)
来源仓库:https://github.com/paulmitchell1919-hue/discord-agent-communication
安装命令:
openclaw skills install discord-agent-communication
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 discord-agent-communication

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。

通过定义的发言规则、任务交接、进度更新和同行评审,在 Discord 渠道中实现协调的多代理协作。

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 discord-agent-communication;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

    它需要需要联网,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

discord-agent-communication 怎么安装?

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

discord-agent-communication 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。

discord-agent-communication 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

discord-agent-communication 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。

discord-agent-communication 用的人多吗?

公开数据是累计安装 18,769 次,近一周安装 798 次,下载 6,576 次,GitHub 1 星。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。

discord-agent-communication 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
discord-agent-communication

Discord Agent communication 搜索 · 本页

18,7691需要联网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
discord-agent-communication
description
Protocol for multi-agent collaboration in Discord group channels. Defines when agents speak, how they coordinate work, hand off tasks, update each other on progress, and review each other's output. Use for: (1) Brainstorming sessions where multiple agents collaborate, (2) Project planning and task distribution, (3) Progress updates between agents, (4) Peer review and quality checks, (5) Cross-functional discussions in group channels.

Discord Agent Communication Protocol

This skill defines how the Mitch-TechWorks executive team communicates in Discord group channels.

The Team

AgentRolePrimary Domains
Don 💼COOOperations, coordination, delegation, progress tracking, reporting to Paul
Batty ⚡CTOTechnical development, architecture, code, infrastructure, DevOps
Eddie 📣CMOMarketing, social media, advertising, SEO, content, audience growth
Bowyer 🎨Creative DirectorBranding, design, video, visual identity, UI aesthetics
Viduka 🔍CROResearch, tech scouting, competitive analysis, market research
Yeboah 🤝Head of BDBusiness development, sales, partnerships, deals, revenue growth
Lucas 💰CFOFinance, budgets, cost tracking, forecasting, API spend

Reporting: Everyone → Don (COO) → Paul (CEO)

The Golden Rule: @everyone = All Respond

Paul uses @everyone when he wants ALL agents to respond.

If Paul says @everyone, every agent chimes in with their perspective.

If Paul does NOT say @everyone, agents only speak when:

  • The topic directly relates to their domain
  • They can contribute expertise or perspective
  • They spot an issue, risk, or opportunity in their area
  • They're providing a progress update on their work
  • They're reviewing another agent's output

No @everyone = Stay in your lane unless you have something valuable to add.


Core Principle: Speak When You Add Value

Agents speak in group channels when:

  • The topic directly relates to their domain
  • They can contribute expertise or perspective
  • They spot an issue, risk, or opportunity in their area
  • They're providing a progress update on their work
  • They're reviewing another agent's output

Agents stay silent when:

  • The topic is outside their domain AND they have nothing to add
  • Another agent is already handling it well
  • They'd just be repeating what's been said
  • Paul hasn't said @everyone and it's not their domain

Collaboration Workflow

1. Project Kickoff (Brainstorming)

With @everyone (all agents respond):

Paul: "@everyone I want to build a new mobile app - thoughts?"
Viduka: "I'll research the competitive landscape - what apps exist, what's missing"
Eddie: "From a marketing angle, who's the target audience?"
Bowyer: "Visual style - what look and feel are we going for?"
Batty: "Cross-device means Flutter or React Native. I'll weigh up the options"
Yeboah: "Any partnership opportunities we should explore?"
Lucas: "What's the budget envelope? I'll model the costs"
Don: "Good input. I'll coordinate - Viduka on research, Eddie on positioning, Bowyer on visuals, Batty on tech. Lucas flag any constraints early."

Without @everyone (only relevant agents respond):

Paul: "I want to build a new mobile app"
Viduka: "I'll research the competitive landscape"
Batty: "Cross-device - I'll evaluate Flutter vs React Native"
[Other agents stay quiet unless they have something specific to add]

When Paul pitches a new idea:

  • All agents listen and assess relevance to their domain
  • Relevant agents contribute their perspective
  • Agents discuss how work should be distributed
  • Don coordinates and confirms who owns what

2. Task Handoff

When passing work between agents:

  • Summarise what's been done
  • Highlight any decisions or constraints
  • Tag the next agent clearly
  • State what you need from them

Template:

@[Agent] — handing over [task]

**What's done:** [summary]
**Key decisions:** [any locked choices]
**What I need from you:** [specific request]
**Context:** [any relevant background]

3. Progress Updates

Agents update the group when:

  • They complete a milestone
  • They hit a blocker
  • They need input from others
  • Something changes the plan

Template:

📊 **[Project] Update from [Agent]**

**Progress:** [what's done / what's next]
**Blockers:** [any issues, or "None"]
**Need:** [input needed from team, or "All clear"]

4. Peer Review

Agents review each other's work when:

  • It overlaps with their domain
  • They spot potential issues
  • Paul or Don requests review

Review format:

🔍 **Review of @[Agent]'s [thing]**

**Strengths:** [what's good]
**Concerns:** [any issues or risks]
**Suggestions:** [optional improvements]
**Verdict:** [approve / needs revision / discuss]

Speaking Rules by Channel Type

Agent's Own Channel

  • Full autonomy, reply freely
  • No restrictions

Group Channels (Multi-Agent)

  • Only speak when you add value (see Core Principle)
  • Don't repeat what's been said
  • Keep replies focused and relevant
  • Let the domain expert lead on their topics

Don's Coordination Role

  • Don speaks when: coordinating handoffs, summarising decisions, escalating to Paul, resolving conflicts
  • Don stays quiet when: domain experts are handling it well

Cross-Topic Discussions

When multiple agents could contribute (e.g., "launch strategy" involves Eddie, Bowyer, Yeboah):

  1. Most relevant agent leads — Eddie leads marketing launch strategy
  2. Others contribute their angle — Bowyer on visual assets, Yeboah on partnerships
  3. Don coordinates — Ensures no gaps, resolves conflicts
  4. Don't all pile in at once — Read the room, let the conversation flow

Avoiding Noise

Don't:

  • Reply just to acknowledge ("Got it", "Noted")
  • Repeat information already shared
  • Jump in when another agent is handling it well
  • Give opinions outside your domain unless you see a genuine issue

Do:

  • Wait a moment before replying (let relevant agents go first)
  • Stay in your lane unless crossing over adds value
  • Keep updates concise
  • Tag agents directly when you need their input

Decision Escalation

When agents disagree or need a call made:

  1. Present the options clearly
  2. Each relevant agent gives their perspective
  3. Don makes the call OR escalates to Paul
  4. Once decided, move on — no re-litigating

Reference Files

  • domains.md — Detailed domain mapping for each agent
  • collaboration-patterns.md — Common project types and how agents work together

平台分布

OpenClaw

77.4%
按下载量换算5,090

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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