Token导航 LogoToken导航

Agent Skill · 研究检索

meeting-synthesizer(会议合成器)

meeting-synthesizer 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 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。

meeting-synthesizer 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证为 MIT,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 21,952。

总安装 4,703周安装 194GitHub Stars 55下载 1,536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

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:meeting-synthesizer(会议合成器)
来源仓库:https://github.com/rysweet/amplihack
仓库路径:skills/meeting-synthesizer
安装命令:
npx skills add https://github.com/rysweet/amplihack --skill meeting-synthesizer
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/rysweet/amplihack --skill meeting-synthesizer

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。

meeting-synthesizer 用于查找、检索和筛选相关信息。适合根据关键词、任务场景或来源线索快速定位候选结果。可在 Codex、Claude、Cursor、Gemini CLI 中辅助信息聚合与匹配。安装前需确认权限范围和维护状态,注意是否触发联网或命令执行。

建议结合原始 README 核验具体功能和使用方式。

来源仓库内路径:skills/meeting-synthesizer

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 Codex、Claude、Cursor、Gemini CLI;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 npx skills add https://github.com/rysweet/amplihack --skill meeting-synthesizer;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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

meeting-synthesizer 怎么安装?

在终端执行 npx skills add https://github.com/rysweet/amplihack --skill meeting-synthesizer。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

meeting-synthesizer 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。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。

meeting-synthesizer 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

meeting-synthesizer 是免费的吗?

来源标注的许可证是 MIT。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。

meeting-synthesizer 用的人多吗?

公开数据是累计安装 4,703 次,近一周安装 194 次,下载 1,536 次,GitHub 55 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。

meeting-synthesizer 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

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,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 Codex,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
meeting-synthesizer

会议合成器 · 本页

4,70355只读MIT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

Meeting Synthesizer Skill

Transform raw meeting notes and transcripts into actionable intelligence. This skill extracts critical information from unstructured meeting data and organizes it into structured formats that drive follow-through and accountability.

Overview

The Meeting Synthesizer processes meeting notes or transcripts to produce:

  • Action Items: Clear tasks with owners, due dates, and context
  • Decisions Made: Key decisions with justification and implications
  • Discussion Points: Important topics, debates, and consensus areas
  • Follow-up Tasks: Prioritized task list with next steps
  • Key Insights: Patterns, concerns, and emerging themes

When to Use This Skill

  • After team meetings, standups, or planning sessions
  • When converting transcript/notes into structured action items
  • When you need to quickly extract decisions and assign accountability
  • When synthesizing notes for stakeholders who missed the meeting
  • When creating task lists from meeting discussions

How to Use This Skill

Basic Workflow

  1. Provide the Meeting Content

- Share meeting notes (typed or pasted) - Share a transcript (from Zoom, Teams, or other platform) - Share a voice-to-text transcription

  1. Request Processing

- Ask the skill to synthesize the meeting - Optionally specify which elements to focus on - Optionally provide format preferences (markdown, JSON, etc.)

  1. Review & Act

- Review the extracted information - Assign owners and deadlines - Create tickets or calendar items - Share results with team

Example Prompts

Comprehensive Synthesis
Synthesize this meeting:
[paste meeting notes/transcript]

Extract all action items, decisions, and key points.
Organize by priority and due date.
Focused on Action Items
Extract action items from this meeting:
[paste meeting notes]

For each action item, provide:
- Task description
- Owner/responsible person
- Due date (if mentioned)
- Context/why it matters
- Dependencies
Decision Extraction
What decisions were made in this meeting?
[paste meeting notes]

For each decision, provide:
- What was decided
- Why (reasoning/justification)
- Who made the decision
- Who is affected
- Any dissent or concerns
Priority Ordering
Synthesize this meeting with priority ordering:
[paste meeting notes]

Order action items by urgency and impact.
Identify any blocking items or critical path items.
Flag any decisions that need follow-up or clarification.

Processing Instructions

1. Parse Meeting Content

Input Recognition:

  • Detect format: structured notes, transcript, bullet points, freeform text
  • Identify participants: who attended, who spoke, roles
  • Extract timestamps: when discussed (if available)
  • Recognize context: meeting type, agenda, goals

Normalization:

  • Clean up formatting inconsistencies
  • Expand abbreviations where possible
  • Clarify ambiguous references
  • Group related items

2. Extract Action Items

Identification:

  • Look for explicit assignments: "Alice will...", "We need to..."
  • Identify implicit tasks: decisions requiring implementation
  • Find time-bound items: "by Friday", "next sprint", "ASAP"
  • Recognize dependencies: "after X is done"

Structuring:

Action Item:
  Description: [Clear, specific task]
  Owner: [Person responsible]
  Due: [Date/timeline]
  Context: [Why it matters]
  Dependencies: [What must happen first]
  Priority: [High/Medium/Low]

3. Identify Decisions

Detection:

  • Explicit decisions: "We decided to...", "We will..."
  • Consensus moments: "Everyone agreed...", "Consensus is..."
  • Choices made: "We chose X over Y"
  • Direction changes: "We're shifting from X to Y"

Documentation:

Decision:
  What: [The decision]
  Why: [Reasoning/justification]
  Decided By: [Who decided]
  Affected: [Who is impacted]
  Dissent: [Any concerns/opposition]
  Implementation: [How will it work]

4. Capture Discussion Points

Extraction:

  • Main topics discussed
  • Debates or different viewpoints
  • Open questions or uncertainties
  • Agreed-upon principles or approaches

Organization:

Discussion Point:
  Topic: [What was discussed]
  Viewpoints: [Different perspectives]
  Consensus: [What was agreed]
  Open Issues: [Unresolved items]

5. Generate Follow-up Tasks

Identification:

  • Clarification needed: "We should confirm with X"
  • Information gathering: "Research options for Y"
  • Scheduling: "Schedule sync with Z"
  • Communication: "Update stakeholders on decision"

Prioritization:

  • Critical: Blocks other work, high impact
  • High: Important, needed soon
  • Medium: Should be done, no immediate urgency
  • Low: Nice to have, lower priority

Output Format

Option 1: Markdown (Default)

# Meeting Synthesis: [Meeting Title/Date]

## Participants

- Alice (Product Lead)
- Bob (Engineering)
- Carol (Design)

## Decisions Made

### Decision 1

- **What**: Adopt Vue 3 for new frontend
- **Why**: Better TypeScript support, smaller bundle size
- **Decided**: By consensus
- **Impact**: Requires training for team members using Vue 2

### Decision 2

- **What**: Delay launch by 2 weeks
- **Why**: Need more time for security review
- **Decided**: Product + Security alignment
- **Impact**: Affects Q4 roadmap planning

## Action Items

### High Priority

- [ ] Complete security audit (Bob, Due: Friday)
- [ ] Prepare Vue 3 migration plan (Carol, Due: Monday)

### Medium Priority

- [ ] Schedule team Vue 3 training (Alice, Due: Next week)

### Low Priority

- [ ] Document decision rationale (Bob, Due: End of sprint)

## Key Discussion Points

- Concerns about timeline slippage
- Team enthusiasm for Vue 3 adoption
- Need for better documentation

## Follow-up Required

- Confirm security audit scope with CTO
- Get client approval for 2-week delay
- Check team training capacity

Option 2: JSON Structure

{
  "meeting": {
    "title": "Q4 Planning",
    "date": "2024-11-08",
    "participants": ["Alice", "Bob", "Carol"]
  },
  "decisions": [
    {
      "id": "D1",
      "description": "Adopt Vue 3",
      "rationale": "Better TS support",
      "decider": "Consensus",
      "impact": "Requires team training"
    }
  ],
  "actionItems": [
    {
      "id": "A1",
      "task": "Complete security audit",
      "owner": "Bob",
      "dueDate": "2024-11-10",
      "priority": "HIGH",
      "context": "Required before launch"
    }
  ],
  "followUp": ["Confirm security audit scope", "Get client approval"]
}

Examples

Example 1: Product Planning Meeting

Input Meeting Notes:

Meeting: Product Planning - Q1 2025
Attendees: Sarah (PM), Mike (Lead Dev), Lisa (Designer)

Topics:
1. User feedback analysis - lots of requests for dark mode
2. Performance issues reported in production
3. Q1 roadmap planning

Decisions:
- Will implement dark mode as primary feature
- Will address performance issues first (performance is blocking)
- Delaying mobile app until Q2

Action items:
Sarah will brief the team on user feedback by tomorrow
Mike will create performance improvement plan by Friday
Lisa will start dark mode design exploration next week
Need to schedule meeting with exec team for roadmap approval

Issues:
- Performance work might delay features
- Team capacity might be tight with both performance + dark mode
- Mobile app delay will disappoint some users

Output:

# Q1 2025 Product Planning Synthesis

## Decisions

1. **Implement dark mode as primary Q1 feature**
   - Why: High user demand from feedback analysis
   - Impact: Becomes focus of design/feature work

2. **Prioritize performance issues before new features**
   - Why: Blocking production users
   - Impact: Requires dedicated engineering time first

3. **Delay mobile app to Q2**
   - Why: Capacity and roadmap constraints
   - Impact: Will disappoint some users, needs communication

## Action Items

### Critical Path

- [ ] Performance improvement plan (Mike, Due: Friday)
  - Context: Blocking roadmap
  - Unblocks: Dark mode feature work

### High Priority

- [ ] Team briefing on user feedback (Sarah, Due: Tomorrow)
- [ ] Dark mode design exploration (Lisa, Due: Next week)

### Follow-up Required

- [ ] Executive team meeting on Q1 roadmap
- [ ] Communication plan for mobile app delay
- [ ] Capacity planning for parallel work streams

## Key Concerns

- Tight team capacity with performance + dark mode
- Mobile app delay needs stakeholder management
- Performance work timeline might slip

Example 2: Engineering Standup

Input Transcript:

Sarah: OK team, quick standup. Mike, how's the API work?

Mike: Good progress. Got the authentication endpoints done.
Got the rate limiting working. Should be done by EOW.

Sarah: Great. Any blockers?

Mike: Need DB schema review from Carol. That's blocking
the user service implementation.

Carol: Oh, I can look at that this afternoon. Will send feedback.

Sarah: Perfect. Lisa, how's the frontend?

Lisa: Integrated the new modal component. Ran into a
responsive design issue on mobile that I'm debugging.
Should have fixed by tomorrow.

Sarah: OK, don't spend too much time on that if it's complex.
We can handle edge cases later.

Mike: Also, we should probably do a load test before shipping.

Sarah: Agreed. Let's add that to the list. Carol, can you
help set that up?

Carol: Sure, I can coordinate with DevOps.

Sarah: Good. Let's reconvene Thursday to check in.

Output:

# Engineering Standup - 2024-11-08

## Current Status

- **Mike**: API auth endpoints complete, rate limiting working
  - On track: EOW completion
  - Blocker: Waiting on DB schema review

- **Lisa**: Modal component integrated, debugging responsive issue
  - On track: Mobile fix by tomorrow
  - Note: Don't over-engineer edge cases

- **Carol**: Available for schema review, coordinating load testing

## Action Items

### High Priority

- [ ] DB schema review (Carol, Due: Today afternoon)
  - Unblocks: Mike's user service implementation

- [ ] Mobile responsive debugging (Lisa, Due: Tomorrow)
  - Context: Modal component issue

### Medium Priority

- [ ] Coordinate load testing setup (Carol + DevOps, Due: This week)
  - Context: Required before shipping

## Decisions

- Load testing is required before deployment
- Don't over-engineer mobile edge cases yet

## Next Check-in

- Thursday team standup for progress review

Processing Tips

Handle Ambiguity

  • If owner unclear: Note "TBD" and flag for clarification
  • If timeline unclear: Use phrases like "asap", "this week", "unclear timeline"
  • If decision unclear: List as "discussion item" vs "decision"

Identify Hidden Items

  • "We should..." often means someone needs to do it
  • "Someone will need to..." is an implicit action item
  • "We need to verify..." is a follow-up task
  • "Can you..." is an action item for the person being asked

Prioritize Effectively

  • Critical: Blocks other work
  • High: Time-sensitive or important
  • Medium: Should be done soon
  • Low: Nice to have

Red Flags

  • Vague ownership ("we'll figure it out")
  • No timeline mentioned
  • Conflicting decisions
  • Unresolved concerns
  • Unclear next steps

Best Practices

Input Quality

  • More detail is better (full transcript > vague summary)
  • Include speaker names if available
  • Preserve disagreements and concerns
  • Note time-sensitive items

Output Quality

  • Be specific: "review code" not clear, "review API endpoint authentication code" is clear
  • Include context: Why matters, not just what to do
  • Flag dependencies: What must happen first
  • Highlight urgency: Critical path items

Follow-Through

  • Share results with team members assigned
  • Create tickets/calendar items for action items
  • Schedule follow-up for open items
  • Update stakeholders on decisions

Common Patterns

Pattern 1: Decision with Cascading Action Items

Decision: Use PostgreSQL instead of MongoDB
  ↓
Action Items:
- Migrate existing data (Owner: DB team)
- Update ORM code (Owner: Backend team)
- Update schema documentation (Owner: Docs)
- Train team on SQL (Owner: Tech lead)

Pattern 2: Blocking Dependencies

Task A: Complete security audit (BLOCKING)
  ↓
Task B: Deploy to production (BLOCKED)
  ↓
Task C: Start Q2 roadmap planning (BLOCKED)

Pattern 3: Decision Reversal or Clarification Needed

Stated: "We'll use Vue"
But earlier: "React is better"
Follow-up: Clarify decision rationale and communicate to team

Quality Checklist

After synthesizing a meeting, verify:

  • All action items have clear owners
  • All decisions are documented with rationale
  • Critical items are flagged and prioritized
  • Blocking dependencies are identified
  • Follow-up tasks are clear
  • Any concerns or dissent are captured
  • Ambiguous items are flagged for clarification
  • Results are ready to share with team

Integration Points

With Task Management

  • Export action items to Jira, GitHub Issues, Asana
  • Create calendar items for deadlines
  • Set up reminders for owners

With Communication

  • Share synthesis with team
  • Send to stakeholders who couldn't attend
  • Create meeting summary email

With Roadmap Planning

  • Decisions impact roadmap/timeline
  • Action items affect capacity planning
  • Follow-up tasks might become next meeting agenda

Limitations & Considerations

  • Cannot identify non-verbal communication (tone, body language)
  • May miss implicit context only meeting attendees understand
  • Requires clear, interpretable notes or transcripts
  • Ambiguous meeting notes produce ambiguous output
  • Cannot follow up on decisions that weren't explicit

See Also

  • Action item tracking and management
  • Decision log and decision history
  • Meeting agenda templates
  • Post-meeting communication templates

平台分布

Claude Code

26.64%
按下载量换算409

Antigravity

23.55%
按下载量换算362

OpenCode

17.66%
按下载量换算271

Gemini CLI

11.77%
按下载量换算181

Codex

6.69%
按下载量换算103

windsurf

3.33%
按下载量换算51

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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。

继续浏览同类 Skills