agentsclimarketplace
Sign in

Tui installer

Skill jajupmochi/agent-harness/skills/tui-installer

Use to install (or just plan) the recommended terminal/TUI stack for driving multiple AI coding agents on Ubuntu — zellij + claude-squad + lazygit + delta. Dry-run by default (prints the plan, installs nothing); --apply asks y/N per tool. When onboarding agent-harness, offer to run --check and ask before installing anything.From its SKILL.md

Install
npx -y skills add jajupmochi/agent-harness --skill tui-installer

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.1 KB, 435 tokens by cl100k_base, as published. Nobody here has run it

tui-installer

Overhaul task 8 (installer half). Turns the recommendation in recommendations/tui-for-agents.md into a runnable, ask-first installer. Nothing installs unless the user opts in.

Use

bash scripts/install-tui.sh --check    # report what's installed vs missing
bash scripts/install-tui.sh            # dry-run: print the install plan, install NOTHING (default)
bash scripts/install-tui.sh --apply    # install missing tools, asking y/N per tool

The recommended stack (see the recommendation doc for why): zellij (multiplexer), claude-squad (agent orchestrator, git-worktree isolation), lazygit (diff review + staging), delta (highlighted diffs). Tools without a clean apt/cargo package (claude-squad, lazygit on some distros) are marked MANUAL: with the upstream install URL rather than a fabricated command.

Onboarding behavior (task 8)

When applying agent-harness to a machine, run --check and ask the user whether to install the missing TUI tools — never install unprompted. Re-check periodically (or when the recommendation doc changes) and ask again if a better tool appears; keep the tool list in sync with recommendations/tui-for-agents.md.

Why dry-run default + fake-install seam

Installing system tools is side-effectful, so the default only plans. The detection is testable without touching the system via TUI_FAKE_INSTALLED="zellij delta" (used by test_install_tui.sh, 11/11). The actual package installs are the only non-deterministic part and stay behind --apply + a y/N prompt.

Status: v0.1 (check / dry-run plan / apply-with-prompt, tested). Planned: keep the tool table generated from the recommendation doc, and a "notify when a better tool ships" tie-in with agent-update-watcher.

What ships with it: 2 files

4.7 KB alongside SKILL.md, 2 of them executable

scripts/

Keep looking

Skills are one crate of 326,144. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.