TreeMap
A free, open-source disk-space visualizer. Squarified treemap, duplicate hunting, disk-full forecasting, and cleanup that only ever moves files to the Trash.
翻譯準備中
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證A local, history-backed work board where users and AI agents collaborate.
翻譯準備中
取自儲存庫 README
正在載入完整 README。
檢視完整 README ↗A shared work board where users and AI agents keep work moving together.
AI agents register work for themselves and users. Users update tasks and answer decisions.
日本語 · Usage guide · Contributing · GitHub repository
When an agent works for a long time, you need to see what is moving, what is waiting, and what response will unblock it. Agent Board gives users and agents one shared, history-backed board.
flowchart LR
A["AI agent"] -- "agboard CLI + hooks" --> D[("Local shared folder")]
U["User"] -- "Agent Board UI" --> D
D -- "local server" --> B["Browser"]
1. One board for people and AI agents. Track work in Todo, Doing, and Done, with Running and Pending states showing what the agent is doing. Answer questions directly on a task from your PC or phone, and keep decisions and status changes in its history.
2. Keep one agent session moving through tasks. Order tasks within an epic and let agboard progress return the work that can advance, together with the paths to its plans and tasks. An agent can explicitly listen for the next eligible task with agboard --project "My project" wait --until task --idle-timeout 300 --json; it processes one task, then listens again in the same session. Existing Todo tasks and idle AI tasks already in Doing are included, and the command exits normally after the chosen idle interval. Folder-based project selection and hooks give agents the right context. When every acceptance condition is verified, eligible AI tasks move to Done automatically; completed work stays read-only for agents.
3. Local Markdown for plans and results. Keep epic details in EPIC.md and tasks in TASK.md, with images alongside them. A task's plan and verification results live in the TASK.md body, with acceptance conditions and history in its frontmatter; task metadata such as its current status lives in task.json. Each history entry has just a timestamp (at) and content (reason). Agents record meaningful work progress automatically, and UI or CLI operations add their own history. You may supply a reason or let Agent Board generate one from the current state. Everything stays in your local data area, with no account or hosted service required.
See the usage guide for editing, attachments, archiving, and settings.
Build from source with Git, Rust (Cargo with edition 2024 support), Node.js, and pnpm:
git clone https://github.com/FooQoo/agent-board.git
cd agent-board
pnpm --dir web install
pnpm --dir web build
cargo install --path . --locked --force
Keep ~/.cargo/bin (or $CARGO_HOME/bin) on your PATH. Build the web UI before installing: agboard embeds it. See the usage guide for updates and installation checks.
agboard project create "My project"
agboard serve --port 7420
Open http://127.0.0.1:7420/, select your project, and add a task. Most board commands use --project NAME; agboard progress and hooks resolve the project from the current folder unless explicitly overridden.
To check progress and answer questions from your phone, open Settings → Mobile connection and connect with your ngrok authtoken. The phone has to enter a 4-digit passcode that the board shows on your computer and that changes each time the connection starts; the full board stays on your computer. You can also start it with NGROK_AUTHTOKEN set and agboard serve --monitor ngrok, which prints the passcode. A 4-digit passcode is weaker than an account sign-in, so keep the address private; the mobile monitor guide explains the limits.