Agent skill

channel-terminal

Run and operate a local terminal/TUI channel for LionClaw using the kernel channel bridge APIs.

Stars 1
Forks 0

Install this agent skill to your Project

npx add-skill https://github.com/moshthepitt/lionclaw/tree/main/skills/channel-terminal

SKILL.md

Terminal Channel Worker

Use this skill when the user wants a real terminal/TUI channel instead of the direct CLI path.

What you run:

  1. attach a terminal to LionClaw,
  2. type messages as a local peer,
  3. see answer, reasoning, status, and error lanes in separate regions.

Under the hood, the worker:

  1. restores the latest interactive session snapshot for (channel_id, peer_id),
  2. posts inbound text to /v0/channels/inbound pinned to the chosen session_id,
  3. long-polls /v0/channels/stream/pull,
  4. renders typed outbound events in separate terminal regions,
  5. starts recovery actions through /v0/sessions/action,
  6. advances its consumer cursor through /v0/channels/stream/ack.

Prerequisites

  • python3
  • uv
  • A configured LionClaw runtime if channel attach needs to start LionClaw for you

Setup

  1. Register the skill and channel:
bash
lionclaw skill add skills/channel-terminal --alias terminal
lionclaw channel add terminal --launch interactive
  1. Attach your terminal to the interactive channel:
bash
lionclaw channel attach terminal

If LionClaw is not already running, channel attach will start it using the default runtime, or the runtime passed with --runtime.

For a fresh manual test home in one command:

bash
./scripts/bootstrap-terminal-test.sh /tmp/lionclaw-terminal-e2e

That helper gives a fresh test home its own loopback bind, so it does not inherit state from another LionClaw daemon already running elsewhere on your machine.

For repeated manual testing, you can use the repo helper:

bash
./scripts/attach-terminal-test.sh /tmp/lionclaw-terminal-e2e

Notes

  • LionClaw enforces pairing. The TUI shows pending pairing state and the exact approval command to run through lionclaw channel pairing approve ....
  • The worker defaults to channel_id=terminal and peer_id=$USER (falling back to $USERNAME or local-user).
  • channel attach restores the latest history_policy=interactive session for that peer, renders the last 12 durable turns into the transcript, and resumes a still-running answer stream from the last durable checkpoint when one exists.
  • The TUI echoes your message immediately, shows a local pending state right away, and then follows live queue.* / runtime.* stream events for the active turn.
  • Slash commands are built into the TUI:
    • /continue
    • /retry
    • /reset
    • /quit
    • /exit
  • The Transcript pane is durable session history plus live answer deltas.
  • The Thinking pane is live-only. It does not replay historical reasoning on attach.
  • Runtime selection normally comes from the running LionClaw service or lionclaw channel attach <id> --runtime ... when attach needs to start it. Attach only reuses a daemon when that daemon belongs to the same LIONCLAW_HOME.
  • A shell debug harness is available in scripts/debug-worker.sh.

Expand your agent's capabilities with these related and highly-rated skills.

Didn't find tool you were looking for?

Be as detailed as possible for better results