term-cli

Manage detached tmux sessions for interactive terminal applications.

99|9|Updated Feb 1, 2026
One-click install
npx skills add https://github.com/EliasOenal/term-cli --skill term-cli
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: term-cli
Source: https://github.com/EliasOenal/term-cli/tree/main/skills/term-cli
Command: npx skills add https://github.com/EliasOenal/term-cli --skill term-cli

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

Interactive terminal programs (servers, REPLs, debuggers, TUIs, SSH prompts, installers) block your agent when run directly in a terminal.

Core Features & Use Cases

  • Start and manage detached tmux sessions so long-lived tasks keep running while you continue working.
  • Send input and capture state using commands like run, wait, send-text/send-key, capture, and status to safely interact with prompts and screen output.
  • Hand off to humans only when needed for passwords and MFA via request/request-wait flows that pause the agent until a human completes the step.
  • Use cases: running npm dev servers, debugging with pdb/gdb, interacting with SSH password/MFA prompts, operating TUIs (vim/htop/less), and performing in-terminal file transfers.

Quick Start

Ask an AI agent to start a detached session, run an interactive command in the background, wait for a prompt, capture the latest output, and then stop the session when done.

Frequently Asked Questions about term-cli

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run an interactive terminal like a dev server without blocking my AI agent?

Interactive terminals like dev servers block AI agents unless run in detached tmux-backed sessions. This approach keeps long-lived tasks running in the background while allowing the agent to continue working.

How does an AI agent handle SSH password and MFA prompts during terminal automation?

SSH password and MFA prompts are handled via human handoff requests that pause the agent. The workflow waits while a human completes the authentication step, then resumes automated terminal control.

Can I send keystrokes and capture output from a detached tmux session programmatically?

Detached tmux sessions support deterministic input injection using send-text, send-key, send-stdin, and send-mouse commands. Output capture modes include plain, annotated, raw, and scrollback options for robust screen state retrieval.

Is it possible to control TUI applications like vim or htop through an automated agent workflow?

TUI applications like vim and htop can be controlled by starting a detached session and injecting specific keystrokes or mouse inputs. The agent captures the resulting screen output to verify the application state.

What is the best way to interact with debuggers like pdb or gdb without hanging the terminal?

Debuggers like pdb and gdb hang the terminal when run directly, but running them in a detached tmux session prevents blocking. The agent sends commands and captures output to step through code without freezing.

Are there limitations when locking interactive terminal sessions for read-only agent observation?

Session locking enforces read-only behavior, preventing the agent from injecting unintended input into interactive terminals. This safety constraint ensures the agent can monitor output without accidentally altering the state of the running process.