con-terminal skill
This skill teaches Claude Code and Codex to share the con terminal: inspect panes, run commands in view and coordinate workers
- 653 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The agent runs commands in open terminal sessions, SSH ones included
- Automatic resume and context handoff share working folders between agents
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/con-terminalDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The agent runs commands in open terminal sessions, SSH ones included
- Automatic resume and context handoff share working folders between agents
This is third-party code. Review the repository files before installing.
What it does
con is a GPU-accelerated terminal with a built-in AI harness for SSH and tmux work. The `con-terminal` skill from the repository connects coding agents to it so Claude Code, Codex and other CLI agents share one terminal workspace. The agent inspects the existing workspace, runs commands where you can see them and does not take the terminal over. Work goes through `con-cli`: a pane sets the visible layout and a surface is a separate terminal session. Agents address workers by surface ID, read fresh OSC 7501 reports and verify results before declaring success. Silence is never treated as completion, and programs without status reports still work. README scenarios: parallel implementation with a separate worktree per writer, one agent reviewing while another runs checks, resuming Claude Code and Codex conversations after a restart and passing context to another agent (Agent Handoff on macOS). The skill installs with `npx skills add nowledge-co/con-terminal --skill con-terminal` when con and con-cli are already installed.
Who it is for. Developers running several agents in a terminal over SSH and tmux who want to see and coordinate them in one window.
Good fit when
- Several agents write code in parallel and each needs its own worktree and pane
- One agent reviews while another runs checks, and the reports stay in terminal sessions
- Claude Code or Codex was restarted and the conversations should reopen where they were
- A task has to move to another agent together with its context on macOS
- An agent should run commands in view without hijacking your terminal
Not a fit when
- You do not use the con terminal: the skill is built around it and its `con-cli`
- You need a stable release: the terminal itself is in active beta, an early beta on Windows and a preview on Linux
- The agent should work invisibly without visible panes: the skill builds the workflow around visible sessions
Example request
Start two workers in the con terminal: one implements the task in a separate worktree and the other runs review and checksLimitations
The skill works only together with the con terminal and its `con-cli`, which must be installed first: `brew install --cask nowledge-co/tap/con-beta` or the `install.sh` script. The terminal is in active beta: macOS is fully supported, Windows is an early beta and Linux a preview. Automatic resume of Claude Code and Codex conversations is described for macOS and zsh and can be turned off in Settings. Context handoff between agents shares one worktree, so writers need coordinating. The skill installs through `npx skills add`, a third-party skills manager.
How to disable. Delete the skill from the agent's skills folder, for example `~/.claude/skills/con-terminal`. `brew uninstall --cask con-beta` removes the terminal, and automatic resume is turned off in Settings.
FAQ
How do I install the con-terminal skill?
Install the con terminal with `con-cli`, then run `npx skills add nowledge-co/con-terminal --skill con-terminal`. The skill lands in the skill folders of supported agents.
How do I install the con terminal itself?
On macOS with `brew install --cask nowledge-co/tap/con-beta` or the `curl -fsSL https://con-releases.nowledge.co/install.sh | sh` script. On Linux the same script puts `con` and `con-cli` into `~/.local/bin`.
What is the difference between a pane and a surface?
A pane is visible layout and a surface is a specific terminal session. The skill addresses workers by surface ID.
How do I continue a conversation after a restart?
Normal Claude Code and Codex launches on macOS in zsh link to surfaces automatically and reopen the same conversation after a restart. It is on by default and can be turned off in Settings.
What about an agent with no status reports?
Nothing special: such programs still work. The skill does not treat silence as completion and verifies results from the output.
Discussion
No questions or reviews yet.
Related
SkillsConduit runs a team of agents in the terminal: tabs with Codex, Claude Code, Gemini CLI and OpenCode over git working copies.
A skill that teaches an agent inside the Kaku terminal to open panes, type into nearby windows and wait for another agent.
Over 1,100 ready-made skills for Claude Code, Codex, Gemini CLI and Cursor, plus an MCP server that picks a skill for the task
A macOS terminal with the termio CLI and a skill: agents spawn sibling sessions and you see which one needs your answer
The tty7 skill teaches Claude Code and Codex to drive tty7 terminal panes: run commands in a real PTY, read the screen and hand work to another agent
The author's personal set: shared agent rules and dozens of skills for GitHub, the browser, macOS, messaging and releases, plus a sync script
README in short
The README presents con as a terminal with an AI harness for people who want a serious terminal first. The section on working with agents explains the con-terminal skill, the pane and surface concepts and scenarios for parallel work, review and resume. Platform status, quick keys, bottom bar modes and installation for macOS, Linux and AUR follow.