Token导航 LogoToken导航

Agent Skill · 开发

flussonic-media-server(Flussonic 媒体服务器)

flussonic-media-server 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

flussonic-media-server 是开发类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描有提醒项,按安装量在开发类的 21,746 个 Skill 里排第 20,764。

总安装 212周安装 9GitHub Stars 公开资料未说明下载 74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
unknown

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:flussonic-media-server(Flussonic 媒体服务器)
来源仓库:https://github.com/a1500llc/flussonic-skill
仓库路径:skills/flussonic-media-server
安装命令:
npx skills add https://github.com/a1500llc/flussonic-skill --skill flussonic-media-server
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 npx skills 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。

skills.shnpx skills
npx skills add https://github.com/a1500llc/flussonic-skill --skill flussonic-media-server

这个 Skill 是做什么的

开发类 Skill 面向写代码、改代码和排查问题,一般会读写仓库文件,部分还会执行构建或测试命令。

用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在协作流程中整理变更。支持围绕仓库状态、代码变更或协作事项进行梳理,适用于 Codex、Claude 等宿主环境。可结合来源仓库和 README 核验具体用法,建议先确认权限与维护状态。

安装方式:github,命令:npx skills add https://github.com/a1500llc/flussonic-skill --skill flussonic-media-server。注意可能触发联网、命令执行或文件读写,使用前请评估风险。

来源仓库内路径:skills/flussonic-media-server

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/a1500llc/flussonic-skill --skill flussonic-med…;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

    它需要接触敏感数据,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

flussonic-media-server 怎么安装?

在终端执行 npx skills add https://github.com/a1500llc/flussonic-skill --skill flussonic-media-server。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

flussonic-media-server 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。

flussonic-media-server 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

flussonic-media-server 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

flussonic-media-server 用的人多吗?

公开数据是累计安装 212 次,近一周安装 9 次,下载 74 次。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。

flussonic-media-server 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

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。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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。
  • 需要处理开发类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

开发类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
skill-vetter

技能审查员 · OpenClaw

8,779,018986操作浏览器MIT-0通过
ontology

本体论 · OpenClaw

6,636,960576只读MIT-0通过
entra-app-registration

内部应用程序注册 · Codex、Claude

6,631,872764接触敏感数据未明确通过
entra-app-registration

内部应用程序注册 · Codex、Claude

3,757,680201接触敏感数据未明确通过
nano-banana-pro

纳米香蕉专业版 · OpenClaw

3,526,898364接触敏感数据MIT-0有提醒
flussonic-media-server

Flussonic 媒体服务器 · 本页

212—接触敏感数据未明确有提醒

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

Flussonic Media Server Skill

Provide expert guidance on Flussonic Media Server — a professional software for video streaming, recording, and delivery. Help users configure, troubleshoot, and optimize their Flussonic deployments.

Architecture Overview

Flussonic Media Server is an all-in-one streaming server that handles:

  • Ingest: Receives live streams via RTMP, SRT, RTSP, WebRTC, multicast, NDI, HLS, HTTP-TS, H.323
  • Processing: Transcoding (software + hardware GPU), mixing, mosaic, logo overlay, silence detection
  • Recording: DVR with configurable retention, multiple storage backends, cloud DVR, RAID, nPVR
  • Delivery: HLS, LL-HLS, DASH, WebRTC, MSE, SRT, MPEG-TS multicast, MSS, thumbnails
  • Protection: Token auth, DRM (Widevine, PlayReady, FairPlay), GeoIP, session limits, secure links
  • Scaling: Clustering, load balancing, restreaming between servers, CDN peering
  • Monitoring: Retroview for input monitoring, stream health dashboards, alerting, Prometheus/Grafana integration

Key Configuration Model

Flussonic uses a text config file (/etc/flussonic/flussonic.conf) or the web UI. Configuration objects:

  • stream — A named live/VOD stream with input sources, processing, and output settings
  • template — Reusable configuration applied to streams matching a pattern
  • dvr — Named global DVR storage config (must be defined before streams reference it)
  • cluster — Server clustering configuration for redundancy and scaling

Basic config pattern (DVR requires a global dvr block, then streams reference it with @name):

# Global DVR config (define storage location)
dvr storage1 {
  root /mnt/storage;
}

stream channel1 {
  input udp://239.0.0.1:1234;
  input tshttp://backup-server/channel1/mpegts backup;
  transcoder vb=4000k size=1920x1080 ab=128k;
  dvr @storage1 7d;
  push rtmp://cdn.example.com/live/channel1;
}

API Overview

Flussonic exposes two APIs:

  1. Control API (/streamer/api/v3/) — Manage configuration: streams, DVR, templates, sessions, events. Uses Basic auth (edit_auth/view_auth credentials).
  2. Streaming API — Media delivery endpoints: HLS (/stream/index.m3u8), DASH (/stream/index.mpd), WebRTC, thumbnails, etc.

Important API patterns:

  • Stream creation/update uses PUT (upsert): PUT /streamer/api/v3/streams/{name}
  • JSON body uses "inputs": [{"url": "..."}] (array of objects), NOT "input": "string"
  • Monitoring: /streamer/api/v3/monitoring/liveness and /streamer/api/v3/monitoring/readiness
  • Stats: /streamer/api/v3/config/stats

When to Read Reference Files

This skill has detailed reference guides. Read them based on the user's question:

User asks about...Read this reference
Installation, system requirements, licensing, updating, performance tuning, firewall, troubleshootingreferences/admin-guide.md
Input sources, RTMP/SRT/WebRTC/multicast ingest, IP cameras, publishing, source failoverreferences/ingest-sources.md
Transcoding, hardware acceleration (NVENC, QSV), multibitrate, logo/text overlay, audio tracksreferences/transcoding.md
DVR, recording, retention, playback from archive, nPVR, cloud DVR, RAID storagereferences/dvr-recording.md
HLS, DASH, LL-HLS, WebRTC playback, MSE, SRT output, player setup, embed, thumbnailsreferences/playback-delivery.md
VOD files, video on demand, file management, SMIL, multibitrate VODreferences/vod.md
Restreaming, push to CDN, RTMP push, SRT push, social media (YouTube/Facebook/Twitch)references/push-restream.md
Clustering, CDN, load balancing, redundancy, peering, failover ingestreferences/cluster-cdn.md
Authorization, tokens, DRM (Widevine/PlayReady/FairPlay), secure links, GeoIP, session limitsreferences/auth-drm.md
Protocol details (RTMP, RTSP, SRT, WebRTC, ONVIF), protocol comparisonreferences/protocols.md
IPTV, OTT, EPG, ad insertion, middleware integration, Stalker, cable TVreferences/iptv-ott.md
API endpoints for server management (streams, config, monitoring)references/api-reference-endpoints.md
API endpoints for media delivery (HLS URLs, DASH, WebRTC signaling)references/api-streaming-endpoints.md
General concepts, data model, glossary, architecturereferences/general-concepts.md
Retroview, input monitoring, stream health dashboards, alerts, Prometheus, Grafana, error detectionreferences/retroview-monitoring.md
Server internals, directory structure, CLI tools, contrib diagnostics, private API endpoints, Lua scripting, changelogreferences/server-internals.md

Read only the relevant reference(s) — don't load all files at once. For complex questions, read 2-3 related references.

Common Workflows

Live Ingest → Transcode → DVR → Restream to CDN

This is the most common workflow. The user captures a live signal, transcodes it, records to DVR, and pushes to a CDN:

stream live_channel {
  input udp://239.0.0.1:1234;

  # Transcode to multibitrate (multiple vb= on one line = multiple video tracks)
  transcoder vb=4000k size=1920x1080 vb=2000k size=1280x720 vb=800k size=640x360 ab=128k;

  # Record with 7 days retention (references global dvr block)
  dvr @storage1 7d;

  # Enable output protocols
  protocols dash hls;

  # Push to CDN
  push rtmp://cdn.example.com/live/channel;
}

Source Failover

Configure multiple sources with automatic failover (priority by order, first = primary):

stream resilient_channel {
  input srt://primary-encoder:9000 streamid="#!::m=request,r=resilient_channel";
  input rtmp://backup-encoder/live/stream;
  input fake://fake;  # Test pattern if all sources fail
}

Hardware Transcoding (NVIDIA)

stream gpu_channel {
  input rtsp://camera:554/stream;
  transcoder hw=nvenc deviceid=0 vb=5000k size=1920x1080 deinterlace=true deinterlace_rate=frame vb=2500k size=1280x720 vb=1000k size=640x360 ab=128k;
}

Note: Intel QSV was deprecated in v25.09 and removed from the transcoder. Use NVIDIA NVENC for hardware acceleration.

Cluster Setup

# All peers share the same cluster_key
# On all servers:
cluster_key mySecretKey;
peer server1.example.com;
peer server2.example.com;

# Stream with cluster_ingest (auto-distributes capture across peers)
stream channel1 {
  input rtsp://camera:554/stream;
  cluster_ingest;
}

# Or pull from another peer:
stream channel1 {
  input m4f://origin-server/channel1;
}

Input Monitoring with Retroview

Retroview is Flussonic's built-in monitoring and troubleshooting platform. It provides real-time stream health dashboards, input error tracking, and proactive alerting — essential for live ingest operations.

Key monitoring capabilities:

  • Input errors: lost packets, TS continuity counter errors, scrambled packets, broken payloads, HTTP errors
  • Server metrics: CPU, scheduler utilization, memory, disk I/O, GPU usage, network bandwidth
  • Alerts: Configurable thresholds for stream failures, mass input loss, transcoding overload, DVR storage
  • Integration: Prometheus metrics export, Grafana dashboards, webhooks (Slack, Telegram, PagerDuty)

For monitoring a live ingest + transcode + CDN restream workflow, the critical metrics to watch are:

  1. Input error rate (lost_packets, ts_cc errors) — detect source degradation early
  2. Scheduler utilization — ensure transcoding isn't overloading the CPU
  3. DVR disk usage — prevent recording failures from full storage
  4. Output push status — detect CDN delivery failures

Read references/retroview-monitoring.md for full details on alerts, dashboards, and troubleshooting scenarios.

DRM Protection (CPIX)

Flussonic uses the CPIX standard for multi-DRM. Config uses a single drm cpix directive with a keyserver:

stream protected_channel {
  input udp://239.0.0.1:1234;
  drm cpix keyserver=http://drm-server/api/drm/cpix?client=myapp&clientToken=SECRET resource_id=protected_channel;
  protocols dash hls;
}

This enables Widevine, PlayReady, and FairPlay simultaneously through the CPIX keyserver.

SRT Ingest & Output

SRT Publish (receive incoming SRT streams):

# Global shared SRT publish port
srt_publish {
  port 9998;
  passphrase 0123456789;
}

stream my_srt_stream {
  input publish://;
}

Per-stream SRT play (output):

stream my_stream {
  input udp://239.0.0.1:1234;
  srt_play {
    port 9300;
  }
}

SRT caller (pull from remote SRT source):

stream srt_pull {
  input srt://remote-server:8888 streamid="#!::m=request,r=stream_name";
}

Response Guidelines

When helping users:

  1. Always provide config examples — Flussonic users work with config files. Show the exact syntax they need.
  2. Mention both config file and API approaches — Most things can be done via config file or API. Mention both when relevant.
  3. Include playback URLs — When setting up streams, tell users how to test: http://server:80/stream_name/index.m3u8 for HLS.
  4. Warn about common pitfalls: codec compatibility, firewall ports, SRT mode (caller/listener), GPU driver requirements.
  5. Reference official docs — Point users to https://flussonic.com/doc/ for the specific topic when appropriate.
  6. Consider the user's scale — Ask if they haven't said: how many streams? what bitrate? what hardware? This matters for performance tuning.
  7. Never directly modify system files — Show config examples for the user to apply themselves. Do not write to /etc/flussonic/flussonic.conf or other system paths, run service commands, or change firewall rules directly. Present the config/commands and let the user execute them on their server. Use the API (PUT /streamer/api/v3/streams/{name}) when the user wants to make changes programmatically.

Important Ports

PortProtocolUsage
80HTTPWeb UI, API, HLS/DASH
443HTTPSSecure web UI and streaming
1935TCPRTMP ingest/publish
554TCPRTSP
8080TCPAPI (alternative)
9000+UDPSRT (configurable)
CustomUDPMulticast ingest

Quick Troubleshooting

  • Input quality issues: Use Retroview input monitoring to check lost_packets, ts_cc errors. See references/retroview-monitoring.md.
  • Stream not starting: Check source URL, firewall, codec support. Use input_monitor in config.
  • High CPU: Enable hardware transcoding (NVENC), reduce resolution/bitrate, check if transcoding is needed. Note: Intel QSV was removed in v25.09.
  • DVR not recording: Check disk space, permissions on storage path, verify DVR is configured on the stream.
  • Playback stuttering: Check bitrate vs. available bandwidth, consider ABR profiles, check server CPU load.
  • API 401: Verify Basic auth credentials match edit_auth/view_auth in flussonic.conf. API base path is /streamer/api/v3/.
  • SRT connection fails: Check firewall UDP ports, verify passphrase (10-79 chars), check streamid format for caller mode.
  • API "authorization_failed": The v3 API requires edit_auth credentials. Use -u user:pass with curl.

平台分布

Codex

35.3%
按下载量换算26

Claude

28.94%
按下载量换算21

Cursor

20.23%
按下载量换算15

Gemini CLI

8.57%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

可疑

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。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