據 AlphaMoat 統計,截至 2026-09,OCFT - OpenClaw File Transfer 累計下載 3,795 次、獲 1 個星標,在全部 5,078 個 AI 技能中排名第 2,784 位,在「AI 智慧體」分類(969 個)中排名第 553 位。
| 平台 | 累計安裝 | 累計下載 | 星標 | 評論 |
|---|---|---|---|---|
| ClawHub ↗ | 118 | 3490 | 1 | - |
| SkillHub ↗ | 118 | 3795 | 1 | — |
各平台的計數口徑不同,統計時間不一致。
技能包含受信任 peer 管理;初始化會生成節點身份和金鑰,分享連線資訊前應確認對方身份,並只向可信物件傳輸檔案。
資料釋出於 2026-10-29
以下是根據技能文件生成的示例,不代表實際執行結果。
2026-09 起統計方法調整:同名技能按作者分別統計,並覆蓋全部技能。圖中虛線為調整前的月份,與之後的資料不直接比較,也不計算跨越調整的增長。
| 月份 | 累計下載 | 環比 | 星標 |
|---|---|---|---|
| 2026-09 | 3795 | — | 1 |
| 2026-08 | 3407 | +2.5% | 1 |
| 2026-07 | 3324 | +6.4% | 1 |
| 2026-06 | 3125 | +4.0% | 1 |
| 2026-05 | 3005 | — | 1 |
以下是作者釋出的原始檔案,按純文字顯示,本站未做修改或稽核;其中的連結不可點選,命令請核實後再執行。抓取於 2026-10-23。
--- name: ocft description: P2P file transfer between AI agents via message channels. Supports chunked transfer, IPFS fallback for large files, and trusted peer management. homepage: https://github.com/stormixus/ocft --- # OCFT - OpenClaw File Transfer Protocol P2P file transfer between AI agents via message channels.
## When to Use Use this skill when: - Transferring files between AI agents over chat channels - Setting up peer-to-peer file sharing with trusted agents - Sending files through Telegram, Discord, Slack, or any text-based channel - Need chunked transfer with integrity verification - Transferring large files using IPFS fallback ## Installation ```bash npm install -g ocft ``` ## Quick Start ```bash # Initialize your node (generates unique ID and secret) ocft init # View your status ocft status # Export your connection info to share with peers ocft export # Add a trusted peer ocft add-peer <nodeId> <secret> --name "Friend" # Or import from URI ocft import ocft://eyJub2RlSWQ... ``` ## CLI Commands ### Core Commands | Command | Description | |---------|-------------| | `ocft init` | Initialize node with unique ID and secret | | `ocft status` | Show node status and configuration | | `ocft show-secret` | Display full secret (careful!) | | `ocft export` | Export connection info as URI | | `ocft import <uri>` | Import peer from ocft:// URI | | `ocft verify <secret>` | Verify if a secret matches yours | ### Peer Management | Command | Description | |---------|-------------| | `ocft add-peer <id> <secret>` | Add a trusted peer | | `ocft remove-peer <id>` | Remove a trusted peer | | `ocft list-peers` | List all trusted peers | | `ocft extend-peer <nodeId> <hours>` | Extend a peer's trust expiry | | `ocft set-ttl <hours>` | Set default secret TTL (0 = no expiry) | ### Configuration | Command | Description | |---------|-------------| | `ocft set-download <dir>` | Set download directory | | `ocft set-max-size <size>` | Set max file size (e.g., `100MB`, `1GB`) | ### IPFS Fallback (for large files) | Command | Description | |---------|-------------| | `ocft ipfs-enable` | Enable IPFS fallback for large files | | `ocft ipfs-disable` | Disable IPFS fallback | | `ocft set-ipfs-provider <provider>` | Set provider: `pinata`, `filebase`, `kubo` | | `ocft set-ipfs-key <key>` | Set IPFS API key | | `ocft set-kubo-url <url>` | Set Kubo node API URL | | `ocft set-ipfs-threshold <size>` | Size threshold for IPFS (e.g., `50MB`) | | `ocft set-ipfs-gateway <url>` | Set custom public IPFS gateway | ## Features - 🔗 **Message-based**: Transfer files through existing chat channels - 📦 **Chunked transfer**: Split large files into small pieces (48KB chunks) - ✅ **Integrity verification**: SHA-256 hash for chunks and files - 🤝 **Request/Accept**: Explicit acceptance or auto-accept policy - 🔒 **Security**: Trusted peer whitelist with secrets - ⏰ **Secret TTL**: Set expiry time for trust relationships - 🔄 **Resume**: Resume interrupted transfers from last chunk - 🌐 **IPFS Fallback**: Use IPFS for files exceeding chunk threshold ## Protocol OCFT messages use a `🔗OCFT:` prefix with Base64-encoded JSON, allowing file transfers over any text-based channel. ## Limitations - Chunk size: 48KB (safe for Base64 in messages) - Default max file size: 100MB (configurable via `set-max-size`) - Designed for text-based channels - IPFS fallback requires provider setup (Pinata, Filebase, or local Kubo) ## Links - **GitHub**: https://github.com/stormixus/ocft - **npm**: https://www.npmjs.com/package/ocft
將開發過程中的經驗、錯誤與修正記錄為 Markdown 檔案,供日後重用並提升為工作區記憶。
用自然語言描述需求,在本機與遠端技能市場尋找技能並取得推薦。
給 AI agent 用的瀏覽器自動化命令列技能。
將使用者明確糾正與自我檢視整理成分層記憶,並依使用情況升級或封存。
在安裝第三方技能之前,按清單審查來源、許可權與危險模式的技能。
將代理記憶與跨技能狀態整理成具型別與約束的知識圖譜。
規劃 AI 代理主動跟進、工作記憶與回顧流程的工作區設定指南。
指導如何建立與更新有效 Claude 技能的寫作指南。
資料月份:2026-09 · 下載/星標/安裝數聚合自公開技能註冊庫(ClawHub、SkillHub),按月更新。口徑說明見資料方法。