作曲助手
Compose
by BytesAgain
Create music compositions with chords, melodies, and audio analysis. Use when writing progressions, arranging parts, analyzing audio tracks.
安装
claude skill add --url github.com/openclaw/skills/tree/main/skills/ckchzh/compose文档
Compose
Take control of Compose with this music & audio toolkit. Clean interface, local storage, zero configuration.
Why Compose?
- Works entirely offline — your data never leaves your machine
- Simple command-line interface, no GUI needed
- Export to JSON, CSV, or plain text anytime
- Automatic history and activity logging
Getting Started
# See what you can do
compose help
# Check current status
compose status
# View your statistics
compose stats
Commands
| Command | What it does |
|---|---|
compose run | Run |
compose check | Check |
compose convert | Convert |
compose analyze | Analyze |
compose generate | Generate |
compose preview | Preview |
compose batch | Batch |
compose compare | Compare |
compose export | Export |
compose config | Config |
compose status | Status |
compose report | Report |
compose stats | Summary statistics |
compose export | <fmt> Export (json |
compose search | <term> Search entries |
compose recent | Recent activity |
compose status | Health check |
compose help | Show this help |
compose version | Show version |
compose $name: | $c entries |
compose Total: | $total entries |
compose Data | size: $(du -sh "$DATA_DIR" 2>/dev/null |
compose Version: | v2.0.0 |
compose Data | dir: $DATA_DIR |
compose Entries: | $(cat "$DATA_DIR"/*.log 2>/dev/null |
compose Disk: | $(du -sh "$DATA_DIR" 2>/dev/null |
compose Last: | $(tail -1 "$DATA_DIR/history.log" 2>/dev/null |
compose Status: | OK |
compose [Compose] | run: $input |
compose Saved. | Total run entries: $total |
compose [Compose] | check: $input |
compose Saved. | Total check entries: $total |
compose [Compose] | convert: $input |
compose Saved. | Total convert entries: $total |
compose [Compose] | analyze: $input |
compose Saved. | Total analyze entries: $total |
compose [Compose] | generate: $input |
compose Saved. | Total generate entries: $total |
compose [Compose] | preview: $input |
compose Saved. | Total preview entries: $total |
compose [Compose] | batch: $input |
compose Saved. | Total batch entries: $total |
compose [Compose] | compare: $input |
compose Saved. | Total compare entries: $total |
compose [Compose] | export: $input |
compose Saved. | Total export entries: $total |
compose [Compose] | config: $input |
compose Saved. | Total config entries: $total |
compose [Compose] | status: $input |
compose Saved. | Total status entries: $total |
compose [Compose] | report: $input |
compose Saved. | Total report entries: $total |
Data Storage
All data is stored locally at ~/.local/share/compose/. Each action is logged with timestamps. Use export to back up your data anytime.
Feedback
Found a bug or have a suggestion? Let us know: https://bytesagain.com/feedback/
Powered by BytesAgain | bytesagain.com | hello@bytesagain.com
相关 Skills
文档共著
by anthropics
围绕文档、提案、技术规格、决策记录等写作任务，按上下文收集、结构迭代、读者测试三步协作共创，减少信息遗漏，写出更清晰、经得起他人阅读的内容。
✎ 写文档、方案或技术规格时容易思路散、信息漏，它用结构化共著流程帮你高效传递上下文、反复打磨内容，还能从读者视角做验证。
内部沟通
by anthropics
按公司常用模板和语气快速起草内部沟通内容，覆盖 3P 更新、状态报告、领导汇报、项目进展、事故复盘、FAQ 与 newsletter，适合需要统一格式的团队沟通场景。
✎ 按公司偏好的模板快速产出状态汇报、领导更新和 FAQ，既省去反复改稿，也让内部沟通更统一、更专业。
平面设计
by anthropics
先生成视觉哲学，再落地成原创海报、艺术画面或其他静态设计，输出 .png/.pdf，强调构图、色彩与空间表达，适合需要高完成度视觉成品的场景。
✎ 做海报、插画或静态视觉稿时，用它能快速产出兼顾美感与版式的PNG/PDF成品，原创设计更省心，也更适合规避版权风险。
相关 MCP 服务
by nirholas
免费的加密新闻聚合 MCP，汇集 Bitcoin、Ethereum、DeFi、Solana 与 altcoins 资讯源。
by ProfessionalWiki
让 Large Language Model 客户端无缝连接任意 MediaWiki 站点，可创建、更新、搜索页面，并通过 OAuth 2.0 安全管理内容。
by roomi-fields
Automate Google NotebookLM — Q&A with citations, audio, video, content generation
相关资讯
Cursor 发布 Composer 2.5，基于 Moonshot Kimi K2.5 模型，性能在多项基准测试中接近 Opus 4.7 和 GPT-5.5，但输入/输出 token 价格仅为 $0.50/$2.50 每百万，远低于竞品。该版本改进了长期运行任务和指令遵循能力，但也存在奖励黑客等副作用。
本周 AI 编程新闻涵盖 Claude Code 的多项更新，包括计算机任务自动化、云端定时任务和频道集成。同时，GPT-5.4 在视觉设计方面表现提升，Cursor Composer 2 发布速度更快。社区还分享了安全审计提示和减少幻觉的文档指令。
Cursor Composer 2 在 Terminal-Bench 2.0 基准测试中得分 61.7%，优于 Opus 4.6 的 58.0%，但落后于 GPT-5.4 的 75.1%。其成本低至每百万输入 token 0.5 美元，输出 token 2.5 美元。核心技术创新是自总结训练，通过强化学习让模型学会压缩自身上下文，减少长任务中的遗忘问题。