How do I install the Wterm skill?

Run npx skillmds@latest add majiayu000/wterm in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Wterm safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: CAUTION, Skill Scanner: PASS. Capability flags: makes network calls, reads secrets. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Wterm?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Wterm free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

majiayu000 (@majiayu000) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Wterm by majiayu000 — Agent Skill

Name: Wterm
Rating: 4.666666666666667 (3 reviews)
Author: majiayu000

Wterm

Web terminal server: manage the wterm-server daemon for browser-based shell access (local or remote via Tailscale). This skill should be used when checking wterm status, listing or creating terminal sessions, killing sessions, or managing the wterm LaunchAgent. Trigger on: wterm, web terminal, remote terminal, remote shell, browser terminal, terminal sessions.

majiayu000 Updated Sep 9, 2026 567 repo stars

File contents

wterm

Manage the wterm-server daemon — a Node.js service providing browser-based shell access via @wterm/react + node-pty. Access locally at https://wterm.localhost (via portless) or remotely from any tailnet device at http://<hostname>:3036.

Commands

Check status

bash ~/.claude/skills/wterm/scripts/status.sh

Reports whether the daemon is running, its PID, port, and active session count.

List sessions

bash ~/.claude/skills/wterm/scripts/list-sessions.sh

Shows all active PTY sessions (id, name, command, PID, client count, last activity).

Create a session

bash ~/.claude/skills/wterm/scripts/create-session.sh           # default: zsh
bash ~/.claude/skills/wterm/scripts/create-session.sh claude     # Claude Code
bash ~/.claude/skills/wterm/scripts/create-session.sh /bin/bash  # explicit shell

Spawns a new PTY session. Default shell is zsh. Pass claude to start a Claude Code session.

Kill a session

bash ~/.claude/skills/wterm/scripts/kill-session.sh <session-id>

Terminates a session by its hex ID (get IDs from list-sessions.sh).

Restart the daemon

bash ~/.claude/skills/wterm/scripts/restart.sh

Restarts the wterm-server LaunchAgent. All active sessions are lost.

Architecture

Daemon: ~/daimones/wterm-server/ — TypeScript, node-pty, ws, Vite React SPA
Port: 3036 (bound to 0.0.0.0 for Tailscale access)
Auth: Bearer token from WTERM_AUTH_TOKEN in ~/.config/env/secrets.env. Auth is fail-closed — server rejects all requests if no token is configured.
LaunchAgent: com.minoan.wterm-server — RunAtLoad, KeepAlive
Logs: ~/.claude/logs/wterm-server.{out,err}.log

Gotchas

Sessions default to zsh, not Claude Code. Pass claude explicitly to create-session.sh.
The daemon must be running on the host machine. If status shows not running, use restart.sh or check launchctl list | grep wterm.
Auth token is required for remote access. If scripts return 401, verify WTERM_AUTH_TOKEN is set in ~/.config/env/secrets.env.
Restarting the daemon kills all active sessions — the scrollback buffer is in-memory only.

---name: wtermdescription: Web terminal server: manage the wterm-server daemon for browser-based shell access (local or remote via Tailscale). This skill should be used when checking wterm status, listing or creating terminal sessions, killing sessions, or managing the wterm LaunchAgent. Trigger on: wterm, web terminal, remote terminal, remote shell, browser terminal, terminal sessions.--- # wterm Manage the wterm-server daemon — a Node.js service providing browser-based shell access via `@wterm/react` + `node-pty`. Access locally at `https://wterm.localhost` (via portless) or remotely from any tailnet device at `http://<hostname>:3036`. ## Commands ### Check status ```bash bash ~/.claude/skills/wterm/scripts/status.sh ``` Reports whether the daemon is running, its PID, port, and active session count. ### List sessions ```bash bash ~/.claude/skills/wterm/scripts/list-sessions.sh ``` Shows all active PTY sessions (id, name, command, PID, client count, last activity). ### Create a session ```bash bash ~/.claude/skills/wterm/scripts/create-session.sh # default: zsh bash ~/.claude/skills/wterm/scripts/create-session.sh claude # Claude Code bash ~/.claude/skills/wterm/scripts/create-session.sh /bin/bash # explicit shell ``` Spawns a new PTY session. Default shell is zsh. Pass `claude` to start a Claude Code session. ### Kill a session ```bash bash ~/.claude/skills/wterm/scripts/kill-session.sh <session-id> ``` Terminates a session by its hex ID (get IDs from `list-sessions.sh`). ### Restart the daemon ```bash bash ~/.claude/skills/wterm/scripts/restart.sh ``` Restarts the wterm-server LaunchAgent. All active sessions are lost. ## Architecture - **Daemon**: `~/daimones/wterm-server/` — TypeScript, node-pty, ws, Vite React SPA - **Port**: 3036 (bound to `0.0.0.0` for Tailscale access) - **Auth**: Bearer token from `WTERM_AUTH_TOKEN` in `~/.config/env/secrets.env`. Auth is fail-closed — server rejects all requests if no token is configured. - **LaunchAgent**: `com.minoan.wterm-server` — RunAtLoad, KeepAlive - **Logs**: `~/.claude/logs/wterm-server.{out,err}.log` ## Gotchas - Sessions default to zsh, not Claude Code. Pass `claude` explicitly to create-session.sh. - The daemon must be running on the host machine. If status shows not running, use `restart.sh` or check `launchctl list | grep wterm`. - Auth token is required for remote access. If scripts return 401, verify `WTERM_AUTH_TOKEN` is set in `~/.config/env/secrets.env`. - Restarting the daemon kills all active sessions — the scrollback buffer is in-memory only.

Frequently asked questions