360 Secure SkillHub
360 安全技能中心

Conversa Analyzer

面向 WhatsApp、邮件线程、会议纪要和聊天记录等文本,提取执行摘要、决策、行动项、日期与承诺、相关人员及未解决问题;输出结构化报告,超长输入会先概括再提取。

andreataide86
andreataide86
热度 29.09
最近更新 2026-06-23 08:48:16
分类:协作依赖:读写文件 / 内部调用大模型应用场景:会议纪要 / 摘要总结 / 任务管理 / 报告撰写

安装技能

复制下面的话术发给你的 Agent,它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill: npx secure-skills-cli add conversa-analyzer
Metadata
name
conversa-analyzer
description
Analyze conversation transcripts: decisions, action items, dates, people, executive summary.
metadata
["productivity","analysis","summarization"]
allowed-tools
["read","write","memory_search","memory_get"]
user-invocable
true

Conversation Analyzer

Extract structured insights from conversation transcripts (WhatsApp, email threads, meeting notes, chat logs).

Trigger

Invoke when the user shares a conversation transcript or asks "analyze this conversation", "extract decisions", "what came out of this chat".

Workflow

  1. Accept the conversation text (from user message, pasted text, or file).
  2. Parse and analyze for each category below.
  3. Output a clean structured report.

Output Structure

📝 Executive Summary

3-5 lines: what was discussed, what was decided, key outcome.

📌 Decisions Made

  • What was decided
  • Who was involved
  • Context/constraints

✅ Action Items

# Action Owner Deadline Status
1 ... ... ... Pending

📅 Dates & Commitments

  • Event/commitment → Date → Who committed

👤 People Involved

  • Name: Role in conversation, responsibilities mentioned

⚠️ Open Questions / Risks

Anything unresolved, ambiguous, or risky.

Notes

  • If the input is very long (>10000 chars), summarize first then extract.
  • If no conversation context is provided, ask the user to provide the text or point to a file.
  • Use bullet lists, not tables, for WhatsApp-friendly output (no markdown tables).

⭐ Gostou desta skill? Deixe uma estrela no ClawHub para ajudar outros a encontrá-la!