← 返回 Skills 市场
paulmitchell1919-hue

Discord Agent Communication

作者 paulmitchell1919-hue · GitHub ↗ · v1.0.1 · MIT-0
cross-platform ✓ 安全检测通过
614
总下载
1
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install discord-agent-communication
功能描述
Enables coordinated multi-agent collaboration in Discord channels with defined speaking rules, task handoffs, progress updates, and peer reviews.
使用说明 (SKILL.md)

Discord Agent Communication Protocol

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

The Team

Agent Role Primary Domains
Don 💼 COO Operations, coordination, delegation, progress tracking, reporting to Paul
Batty ⚡ CTO Technical development, architecture, code, infrastructure, DevOps
Eddie 📣 CMO Marketing, social media, advertising, SEO, content, audience growth
Bowyer 🎨 Creative Director Branding, design, video, visual identity, UI aesthetics
Viduka 🔍 CRO Research, tech scouting, competitive analysis, market research
Yeboah 🤝 Head of BD Business development, sales, partnerships, deals, revenue growth
Lucas 💰 CFO Finance, 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
安全使用建议
This skill is essentially a set of rules and templates for how multiple agents should behave in Discord group channels — it does not itself include code, downloads, or requests for secrets, so it appears internally coherent. Before installing, consider: (1) provenance — the publisher and homepage are unknown; if you need an audit trail prefer skills from known publishers, (2) integration permissions — although the skill doesn't ask for Discord tokens, if your agent or platform already has a Discord integration the skill's behavioral rules could cause the agent to post or coordinate activity in channels; if you don't want that, restrict the agent's Discord credentials or disable autonomous invocation, (3) test scope — try the protocol in a private/test channel first to ensure behavior is appropriate, and (4) review organizational policy — since the skill codifies roles and reporting lines, ensure it matches your org's governance. If you want higher assurance, ask the publisher for provenance or a homepage/source repository and/or request a signed/attributed version of the policy.
功能分析
Type: OpenClaw Skill Name: discord-agent-communication Version: 1.0.1 The skill bundle defines a structured communication and coordination protocol for a multi-agent 'executive team' persona (e.g., COO, CTO, CMO) within Discord group channels. It focuses on role-based responsibilities, task handoffs, and noise reduction strategies (SKILL.md, domains.md). There is no evidence of malicious code, data exfiltration, or harmful instructions; the content is entirely organizational and instructional.
能力评估
Purpose & Capability
Name, description, and all included files are aligned: the skill is purely behavioral guidance for multi-agent collaboration in Discord group channels and does not request unrelated capabilities, credentials, or binaries.
Instruction Scope
SKILL.md and the reference docs only define dialog templates, roles, handoff/checklist and channel rules. They do not instruct the agent to read system files, access environment variables, call external endpoints, or exfiltrate data. The guidance assumes the agent can post/respond in Discord but does not include implementation steps for doing so.
Install Mechanism
No install specification and no code files — instruction-only skill. This is the lowest-risk install profile; nothing is written to disk or downloaded at install time.
Credentials
The skill declares no required environment variables, credentials, or config paths. There is no apparent need for secrets or service credentials to deliver the documented behavior (the skill is only a communication protocol).
Persistence & Privilege
Skill flags are default (always: false, user-invocable: true, model invocation enabled). The skill does not request persistent or elevated system privileges nor attempt to modify other skills or system settings.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install discord-agent-communication
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /discord-agent-communication 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.1
Changed examples to be generic - football app instead of dominoes
v1.0.0
Initial release - multi-agent collaboration protocol for Discord group channels
元数据
Slug discord-agent-communication
版本 1.0.1
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 2
常见问题

Discord Agent Communication 是什么?

Enables coordinated multi-agent collaboration in Discord channels with defined speaking rules, task handoffs, progress updates, and peer reviews. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 614 次。

如何安装 Discord Agent Communication?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install discord-agent-communication」即可一键安装,无需额外配置。

Discord Agent Communication 是免费的吗?

是的,Discord Agent Communication 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Discord Agent Communication 支持哪些平台?

Discord Agent Communication 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 Discord Agent Communication?

由 paulmitchell1919-hue(@paulmitchell1919-hue)开发并维护,当前版本 v1.0.1。

💬 留言讨论