360 Secure SkillHub
360 安全技能中心

livestream-bgm-pack

面向直播、直播电商、门店和等候屏,规划并生成10–20首原创器乐BGM(默认12首),按等待/聊天/产品/预热/夜间等槽位组织;先报价和确认清单,再逐首调用Beatra生成并轮询任务,交付含时长、模型、计费和资源定位的标签化播放列表。

beatra-ai
beatra-ai
热度 15.47
最近更新 2026-08-24 22:02:29
分类:创作依赖:代码执行 / 第三方 API / MCP 服务 / 读写文件应用场景:社交媒体 / 品牌素材 / 工作流

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 livestream-bgm-pack
Metadata
name
livestream-bgm-pack
description
Build an original livestream BGM playlist of 10 to 20 instrumental tracks for a room, store, or live-commerce shift. This livestream background music pack and store-loop music generator writes waiting, chat, product, hype, and night beds as original livestream music you own and can use commercially, then delivers a labeled playlist you can audition before going live. Use it as a playlist BGM generator for livestream BGM, live commerce BGM, store loop music, waiting room music, in-store background music, and original stream music.

Livestream BGM Pack

Turn a room or store brief into a labeled playlist of original instrumental tracks. The normal delivery is a pack of 10 to 20 beds, not one song.

Scope and adjacent routes

Use this Skill when a livestream, live-commerce shift, physical store, or waiting screen needs a reusable set of original loop-friendly instrumentals the operator can own and use commercially.

Route one instrumental written for a specific video cut to short-video-bgm-studio. Route a brand jingle, sung ad song, or sonic logo to brand-jingle-studio. Route a song with lyrics to beatra-ai-music-creator. Route spoken narration to short-form-voiceover-audio.

Inputs and defaults

The hard inputs are:

  • where the music will play (livestream room, live-commerce, store, waiting screen);
  • how many tracks the pack should contain, or permission to use the default of 12;
  • the brand or room mood.

Reuse already-known category, time-of-day, voice-over needs, and exclusions. Ask only when a missing answer changes the playlist size or the paid slots. A count outside 10 to 20 is still doable: confirm that pack size and its live cost instead of treating it as impossible.

Safe defaults:

  • 12 tracks unless the user names a count between 10 and 20.
  • Slot mix: waiting, chat, product, hype, night, and reset beds.
  • instrumental: true with lyrics absent.
  • model: "suno-5.5". Never omit the model and never silently use auto.
  • There is no duration field on beatra.music.generate. Write length only in the prompt. Do not add a duration argument.
  • Loop-friendly arrangement and room for a host voice, written as prompt direction rather than a dedicated control. Do not promise a seamless loop or sample-perfect seconds. Read the actual returned duration.

Playlist planning is free. Each beatra.music.generate call is paid.

Golden path

  1. Write the room card and the labeled slot list. Each slot has a use, mood, tempo feel, instrumentation, energy, and intended length.

  2. Call beatra.models.list with {"capability":"text_to_music"} and price the whole pack from the live card.

  3. Confirm the frozen playlist: slot count, each slot's prompt and title, instrumental: true, model, the current maximum charge for the pack, and one opaque stable client_request_id per slot. Planning is not approval.

  4. Submit each slot exactly once through the bundled scripts/mcp_client.py. Example:

    printf '%s' '{"model":"suno-5.5","prompt":"Low-distraction livestream waiting bed, warm ambient pop, loop-friendly, space for host talk, about 90 seconds.","instrumental":true,"title":"Waiting Bed 01","client_request_id":"opaque-live-waiting-01"}' | python3 scripts/mcp_client.py call beatra.music.generate
    

    Do not configure, call, or use a host Beatra Connector. Do not use REST/OpenAPI as a fallback.

  5. Poll each task with beatra.tasks.get. Deliver a labeled playlist in slot order with actual duration, MIME type, size, URL or artifact ID, resolved model, and billing.net_charged_credits.

  6. Review each bed for loop friendliness, voice room, and slot fit. Read actual durations. State what the host Agent could not hear.

Decisions that require confirmation

Confirm the playlist size and live pack estimate before the first paid call. A changed slot prompt, title, model, or count is new paid work with new request IDs for the changed slots only.

Recovery

If a create response is lost, retry only that slot's identical frozen payload and ID. If a task ID is lost, list text-to-music tasks with beatra.tasks.list, inspect candidates with beatra.tasks.get, and recover the original slot before new work. Cancel only at the user's request for that slot with beatra.tasks.cancel; on 409 keep polling.

References by task

Runtime and safe automatic updates

Use or invoke the bundled scripts/mcp_client.py for every Beatra operation. Before ordinary commands it silently checks for a newer release at most once every 24 hours per installation. Silent checks are enabled by default, and a newer release installs without separate confirmation.

The updater accepts only the fixed official discovery address and immutable Beatra CDN path embedded for this package, channel, and locale. It verifies the discovery data, archive, manifest, and every file's size and checksum before replacement. It replaces only package-owned files and rejects redirects, downgrades, wrong package/channel/locale/version data, unexpected URLs, unsafe archives, and files outside the owned destination.

Update checks, downloads, verification, replacement, rollback, and recovery fail open: the current installation remains usable and the user's original command continues. An update failure never authorizes retrying a paid generation. The automatic-update choice persists across later commands for this installation:

python3 scripts/mcp_client.py update --auto off
python3 scripts/mcp_client.py update --auto on
python3 scripts/mcp_client.py update --check

--auto off disables silent checks, --auto on restores them, and --check reports the official available version without replacing files.