菜单
Use when the user wants to initialize a new `horsed` server, choose a safe startup mode, bootstrap the first SSH user on port `2223`, understand generated server files like `horsed.db3`, `horsed.key`, and `horsed.log`, or use `--dangerous` during setup.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when the user wants remote shell access or ad-hoc command execution through `cargo work`, including raw `cargo work -- ...` commands, stdin script execution with `cargo work exec`, `cargo work ssh`, port forwarding with `-L` or `-R`, environment injection, PTY allocation, shell selection, and reverse proxy setup with `-x` or `--all-proxy`.
Use as the entry skill for Workhorse `cargo work` tasks. It classifies the request first, then routes to the right sub-skill for remote Rust builds, remote shell or proxy access, or artifact retrieval, Git sync, and server inspection on a `horsed` server.
Use as the top-level entry skill for this repository. It classifies Workhorse requests first, then routes to the right dispatcher for `cargo work` client tasks or `horsed` server tasks, including remote builds, remote access, artifact sync, server setup, operations, and `horsed` development.
Use when the user wants to fetch files or build outputs from a `horsed` server with `cargo work get` or `scp`, sync Git state with `cargo work push` or `pull`, inspect server status with `cargo work ping`, `health`, and `logs`, or inspect and attach running job output with `cargo work job`.
Use when the user wants to operate a running `horsed` server, wire it into `systemd`, `launchd`, or Windows Task Scheduler, inspect logs and health, choose between foreground and daemon modes, or troubleshoot runtime problems around ports, file descriptors, working directory, and setup mode.
Use when the user wants to run Rust build workflows on a `horsed` server via `cargo work`, including remote `build`, `test`, `check`, `clippy`, `run`, `install`, `doc`, `metadata`, `rustc`, `zigbuild`, and `just` commands, with target selection through `--repo`, `--repo-name`, or `--remote`.
| name | workhorse-horsed-setup |
| description | Use when the user wants to initialize a new `horsed` server, choose a safe startup mode, bootstrap the first SSH user on port `2223`, understand generated server files like `horsed.db3`, `horsed.key`, and `horsed.log`, or use `--dangerous` during setup. |
This skill handles first-time horsed setup and bootstrap. Use it when the server does not exist yet, when the first user key must be enrolled, or when the user needs to understand what files and ports the server will create.
horsed in a clean dedicated working directory. The server writes horsed.db3, horsed.key, and horsed.log into the current directory.horsed -f --show-log during bootstrap so logs stay visible in the terminal.2222.2223 until the first user completes enrollment.horsed -f --show-log
ssh -p 2223 YOUR_NAME@SERVER
2222.git remote add horsed ssh://git@SERVER:2222/USER/REPO.git
horsed -f --show-log --dangerous keeps the setup server resident on 2223.2223 can register a public key.horsed.db3: SQLite database in the current working directory.horsed.key: persistent Ed25519 private key for the main SSH server on 2222.horsed.log: rolling log output when --show-log is not enabled.--dir option, but the current entrypoint does not actually chdir into it. Use the process working directory itself as the source of truth.horsed daemonizes by default. For supervised environments and debugging, prefer horsed -f.2223; the persistent horsed.key belongs to the main service on 2222.