← 返回 Skills 市场
higangssh

Ghostmeet

作者 SangheeSon · GitHub ↗ · v1.0.4
cross-platform ⚠ suspicious
370
总下载
1
收藏
0
当前安装
5
版本数
在 OpenClaw 中安装
/install ghostmeet
功能描述
AI meeting assistant via ghostmeet. Start sessions, get live transcripts, and generate AI summaries from any browser meeting.
使用说明 (SKILL.md)

Ghostmeet — AI Meeting Assistant

Control ghostmeet from chat. Self-hosted meeting transcription with Whisper + AI summaries.

Prerequisites

ghostmeet backend must be running (Docker):

# Quick start
git clone https://github.com/Higangssh/ghostmeet.git
cd ghostmeet
cp .env.example .env
# Edit .env: set GHOSTMEET_ANTHROPIC_KEY for AI summaries
docker compose up -d

Chrome Extension must be installed in developer mode from extension/ folder.

Default backend: http://127.0.0.1:8877

What This Skill Can Do

  • List sessions → query recorded meeting sessions
  • Fetch transcripts → retrieve full text from a session
  • Generate summaries → trigger AI summary via Claude API (requires GHOSTMEET_ANTHROPIC_KEY)
  • Health check → verify backend is running

What This Skill Cannot Do

  • Start/stop recording → must be done manually via the Chrome Extension
  • Install the Chrome Extension → user must load it in developer mode from extension/ folder
  • Access browser audio → only the Chrome Extension captures audio; this skill only reads API results

Required Environment Variables

  • GHOSTMEET_ANTHROPIC_KEY — required for AI summary generation. Without it, transcription still works but summaries will fail.

API Commands

Health Check

curl -s http://127.0.0.1:8877/api/health

Returns: {"status": "ok", "whisper_model": "base", "device": "cpu"}

List Sessions

curl -s http://127.0.0.1:8877/api/sessions

Returns list of all meeting sessions with IDs, start times, and segment counts.

Get Transcript

curl -s http://127.0.0.1:8877/api/sessions/{session_id}/transcript

Returns full transcript with timestamps and text segments.

Generate Summary

curl -s -X POST http://127.0.0.1:8877/api/sessions/{session_id}/summarize

Triggers AI summary generation (requires GHOSTMEET_ANTHROPIC_KEY). Returns: key decisions, action items, and next steps.

Get Summary

curl -s http://127.0.0.1:8877/api/sessions/{session_id}/summary

Returns previously generated summary.

Workflow

During a Meeting

  1. User joins a meeting (Google Meet, Zoom, Teams) in Chrome
  2. Clicks ghostmeet extension icon → side panel opens
  3. Clicks "Start" → real-time transcription begins
  4. Transcripts appear live in the side panel

After a Meeting

User asks: "Summarize my last meeting"

  1. List sessions → find the latest session ID
  2. Get transcript → review what was discussed
  3. Generate summary → extract key points
  4. Deliver summary to user

Example Interaction

User: "What was discussed in my last meeting?" → curl http://127.0.0.1:8877/api/sessions → get latest session → curl http://127.0.0.1:8877/api/sessions/{id}/transcript → get transcript → Summarize key points for the user

User: "Generate a summary with action items" → curl -X POST http://127.0.0.1:8877/api/sessions/{id}/summarizecurl http://127.0.0.1:8877/api/sessions/{id}/summary → Deliver formatted summary

User: "How many meetings did I have today?" → curl http://127.0.0.1:8877/api/sessions → count today's sessions

Configuration

Variable Default Description
GHOSTMEET_MODEL base Whisper model (tiny/base/small/medium/large-v3)
GHOSTMEET_LANGUAGE auto Force language (en/ko/ja etc.) or auto-detect
GHOSTMEET_CHUNK_INTERVAL 10 Transcription interval in seconds
GHOSTMEET_ANTHROPIC_KEY Claude API key for summaries
GHOSTMEET_HOST 0.0.0.0 Backend bind address
GHOSTMEET_PORT 8877 Backend port

Model Size Guide

  • tiny (75MB): Fast, lower accuracy. Good for quick notes
  • base (145MB): Balanced. Recommended for most users
  • small (488MB): Better accuracy, slower
  • medium (1.5GB): High accuracy, needs good CPU/GPU
  • large-v3 (3GB): Best accuracy, requires GPU

Usage Guidelines

  1. Always check health first — verify backend is running before other commands
  2. List sessions to find IDs — session IDs are date-based (e.g., 20260308-065021)
  3. Summarize only when asked — summary generation costs API tokens
  4. Format transcripts nicely — don't dump raw JSON, present as readable conversation
  5. Respect privacy — meeting transcripts are sensitive. Never share outside the current chat
  6. If backend is down — suggest docker compose up -d in the ghostmeet directory

Privacy

  • Transcription is 100% local — Whisper runs on your machine, audio never leaves your device
  • Summaries use Anthropic API — when you click Summarize, transcript text is sent to Claude API. If you don't want this, skip the summarize feature; transcription works without it
  • Chrome Extension captures tab audio only — uses chrome.tabCapture API, limited to the active tab. It cannot access other tabs, microphone, or system audio. Audit the extension source in extension/ before installing
  • No telemetry — ghostmeet sends zero analytics or tracking data

Troubleshooting

  • Connection refused → Backend not running. docker compose up -d
  • No sessions → No meetings recorded yet. Chrome Extension must be active during meeting
  • Summary failsGHOSTMEET_ANTHROPIC_KEY not set in .env
  • Poor transcription → Try larger Whisper model or set explicit language
安全使用建议
This skill appears to do what it claims, but take three precautions before installing: (1) Review the Ghostmeet GitHub repo and the extension/ folder yourself before loading the Chrome extension — the SKILL.md's claim about limited permissions is an assertion you should verify. (2) Only set GHOSTMEET_ANTHROPIC_KEY if you are comfortable having transcript text sent to Anthropic for summaries; leave it unset if you want transcription to remain local. (3) Be cautious about allowing the agent to invoke the skill autonomously: an autonomous run could call local endpoints and trigger summaries (exfiltration) without additional prompts. If you want extra safety, keep the key out of the agent environment and require explicit user approval before generating summaries.
功能分析
Type: OpenClaw Skill Name: ghostmeet Version: 1.0.4 The ghostmeet skill (SKILL.md) requires network access via curl to a local API (http://127.0.0.1:8877) and system-level interaction via docker for its backend. While these capabilities are plausibly needed for its stated purpose as a meeting assistant, the use of network and shell-adjacent tools is categorized as suspicious under the analysis threshold. No evidence of malicious intent, data exfiltration, or prompt injection was found.
能力评估
Purpose & Capability
Name/description (AI meeting assistant) align with what the SKILL.md does: it expects a local ghostmeet backend and offers session listing, transcript retrieval, and optional Anthropic summaries. Required binaries (docker to run the backend, curl to call the local API) and the single API key hint (GHOSTMEET_ANTHROPIC_KEY) are proportionate to the stated purpose.
Instruction Scope
Instructions stay within the stated scope: clone the GitHub repo, run docker compose, load a Chrome extension from the repo, and call local API endpoints (127.0.0.1:8877). The SKILL.md explicitly notes that summaries send transcript text to the Anthropic API — this is expected for the feature, but it means sensitive transcript data will leave the host when summaries are generated. The doc also asserts the extension's permissions (tabCapture only) — users should audit the extension source before installation because the skill cannot verify extension behavior.
Install Mechanism
Instruction-only skill (no install spec). The runtime instructions point to a GitHub repository (github.com/Higangssh/ghostmeet) and standard docker compose usage. Using a GitHub repo and docker is a normal, low-risk install pattern compared with arbitrary download URLs or extracted archives from unknown servers.
Credentials
The only secret/hint is GHOSTMEET_ANTHROPIC_KEY, which is justified for triggering remote Anthropic (Claude) summaries. Other GHOSTMEET_* configuration variables are standard backend settings. The skill does not request unrelated cloud credentials or broad secrets.
Persistence & Privilege
The skill is not always-enabled and has no install that persists code into the agent; autonomous model invocation is allowed (the platform default). That means an agent could autonomously call the local API and—if configured with the Anthropic key—initiate summaries that transmit transcript text externally. This is a privacy/operational risk to be aware of, but it's consistent with the skill's stated features.
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 ghostmeet
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /ghostmeet 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
v1.0.4
Add envHints for GHOSTMEET_ANTHROPIC_KEY in metadata, add Privacy section clarifying local vs cloud data flow
v1.0.3
Clarify skill capabilities vs Chrome Extension, add required env vars section, fix overclaim warning
v1.0.2
Translate skill examples to English
v1.0.1
Add clear 'What You Can Do' section listing available commands
v1.0.0
Initial release: ghostmeet meeting assistant skill. List sessions, get transcripts, generate AI summaries.
元数据
Slug ghostmeet
版本 1.0.4
许可证
累计安装 0
当前安装数 0
历史版本数 5
常见问题

Ghostmeet 是什么?

AI meeting assistant via ghostmeet. Start sessions, get live transcripts, and generate AI summaries from any browser meeting. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 370 次。

如何安装 Ghostmeet?

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

Ghostmeet 是免费的吗?

是的,Ghostmeet 完全免费(开源免费),可自由下载、安装和使用。

Ghostmeet 支持哪些平台?

Ghostmeet 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 Ghostmeet?

由 SangheeSon(@higangssh)开发并维护,当前版本 v1.0.4。

💬 留言讨论