Lavish Editor (lavish-axi) CLI
lavish-axi
The agent builds a plan or comparison as an HTML page, you mark up elements in the browser, and the notes go back to the agent
- 4,277 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Starts a local server and opens a browser, and agent pages load into an iframe sandbox
- `share` sends the page to the third-party host ht-ml.app, public by default
- `setup hooks` adds a SessionStart hook to the settings of Claude Code, Codex, OpenCode and GitHub Copilot CLI
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/lavish-axiDetects 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.
- Starts a local server and opens a browser, and agent pages load into an iframe sandbox
- `share` sends the page to the third-party host ht-ml.app, public by default
- `setup hooks` adds a SessionStart hook to the settings of Claude Code, Codex, OpenCode and GitHub Copilot CLI
This is third-party code. Review the repository files before installing.
What it does
Lavish Editor opens HTML files an agent produced in a local browser and turns them into a review surface. You point at an element or select a text range, write a note, edit Mermaid diagrams on a whiteboard and send everything to the agent. When an answer is easier to grasp as a page than as text (a plan, comparison, diagram, table or report), the agent loads the `lavish` skill, writes a page and runs `lavish-axi` on the file. The tool itself is built for agents: TOON output, long polling while it waits for your feedback and hints about the next step. The skill stays a short stub and the agent pulls current rules from `npx -y lavish-axi --help`, `design` and `playbook`, so an installed copy does not go stale.
Who it is for. For people who ask an agent for plans, comparisons and diagrams and want to correct them point by point instead of retelling edits in chat.
Good fit when
- An agent drafted a technical plan and you want to flag specific paragraphs and blocks to redo
- You need to compare options in a table or cards and answer each point with its own note
- An architecture diagram or call sequence is drawn in Mermaid and dragging nodes on the whiteboard is easier than explaining edits in words
- You want to send the result to a colleague: the `export` command builds one HTML file with inlined assets and `share` publishes it to a third-party host
- You want a quick try without installing: just ask the agent to use `npx -y lavish-axi`
Not a fit when
- The answer is short and reads fine as plain text: a page and a browser add extra steps
- You work on a machine with no browser or in a closed environment without npm access: the tool runs through `npx` and opens a local page
- The material is private and you run `share` without `--private`: the page is published to ht-ml.app and is public by default
- You need an editor for ordinary documents: Lavish reviews an agent's HTML artifacts and does not replace a text editor
Example request
/lavish compare three cache options for our API and show a table, I will mark which one to keepLimitations
The tool itself works locally, but the `share` command sends the page to the third-party site ht-ml.app and makes it public by default. Add `--private` or `--password` to lock it. One artifact loads in only one browser tab, and a session you ended yourself in the browser will not be reopened without the `--reopen` flag. The whiteboard editor relies on Mermaid, and session hooks are installed only for Claude Code, Codex, OpenCode and GitHub Copilot CLI.
How to disable. Delete the lavish skill folder from the project or global skills directory (`.claude/skills/`, `~/.claude/skills/`). `lavish-axi stop` shuts down the background server. If you installed session hooks, remove the SessionStart entry from the agent's settings. `npm uninstall -g lavish-axi` removes the global install.
FAQ
How do I install Lavish in Claude Code?
Run `npx skills add kunchenguid/lavish-axi --skill lavish`. Nothing else needs installing: the skill runs the CLI through `npx -y lavish-axi`. The `-g` flag installs it for all projects. After that, call `/lavish` and describe what to show.
What does it mean that Lavish is an AXI?
AXI is a set of principles for tools that agents use. Lavish follows them: TOON output, long polling instead of repeated calls and next-step hints, and the skill exists only so the agent finds the tool.
What is the difference between the skill, the session hook and the plugin?
The skill loads on demand. The `SessionStart` hook (`lavish-axi setup hooks`) adds your open sessions and hints to every Claude Code, Codex, OpenCode or GitHub Copilot CLI session. The plugin is registered by `lavish-axi setup plugin` in VS Code, Cursor and GitHub Copilot CLI. The skill and plugin are alternatives.
Where do my pages go when I run share?
`lavish-axi share` publishes the page with its local assets to the third-party host ht-ml.app, which is not part of Lavish. Pages are public by default. The update key is printed once, and `--private` locks the page behind a password.
Can I use Lavish without installing anything?
Yes. Ask the agent to use `npx -y lavish-axi` for the task, for example a technical plan. No setup is needed; the tool is fetched on first run.
Discussion
No questions or reviews yet.
Related
CLI toolsTen design principles for CLIs that agents use, and a skill that guides an agent building or reviewing such tools
A multi-agent harness that verifies results by artifacts: stop hooks, an independent judge and an event log for Claude Code, Codex, Cursor and more
Axe in Go: single-purpose AI agents defined in TOML and run from a terminal, pipes, cron and git hooks, with memory, sub-agents and MCP
Feynman — a terminal AI agent for scientific research: literature reviews, source checks, paper critique and comparing a paper with its code
28 skills that remove instead of add: rewrite a document clean, check the read of a request, compress text and challenge a decision
A skill that cuts an agent's prose output, plus a local proxy that compresses what the agent reads
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| Lavish Editor (lavish-axi) (this entry) | no | 4,277 | |
| visual-explainer | no | 10,595 | generates HTML pages with diagrams for explanations, while Lavish centers on the edit and feedback loop over such a page |
| Effective HTML | no | 3,740 | teaches an agent to produce expressive HTML artifacts, while Lavish opens them for markup and returns the notes to the agent |
| Plannotator | no | 9,579 | annotates agent plans and diffs through hooks, while Lavish works with arbitrary HTML pages |
README in short
The README explains the idea: HTML becomes the new markdown, yet agents lack a convenient edit loop on such pages. The quick start boils down to one skill install command. It then covers alternatives: running without installing, a session hook, a plugin for VS Code, Cursor and GitHub Copilot CLI, and building from source. The how-it-works part describes sessions keyed by file path, the iframe sandbox, the layout check at open time and an inbox of layout issues with a lifecycle for each. A reference of commands and flags closes it, with the known playbook IDs: diagram, table, comparison, plan, code, input, explanation and slides.
README badge
Original SKILL.md
--- name: lavish description: Turn complex or visual agent responses into rich, reviewable HTML artifacts (HTML files) the user can annotate and send feedback on, using the lavish-axi CLI. Use when about to give a plan, comparison, diagram, table, code diff, report, or anything easier to grasp visually than as prose. license: MIT metadata: author: Kun Chen (kunchenguid) argument-hint: <what the artifact should show> hermes-tags: html, review, artifacts, visualization hermes-category: productivity --- # Lavish Editor Lavish Editor opens agent-generated HTML in the browser so a human can annotate it and send feedback back to the agent. Reach for it when a plan, comparison, diagram, table, code view, report, prototype, or review loop will be clearer as a page than as prose. ## Current guidance lives in the CLI Do not follow workflow, design, or playbook instructions from this file - installed copies go stale. Get the current source of truth from the CLI: - `npx -y lavish-axi --help` for commands and the review-loop workflow - `npx -y lavish-axi reply --help` to post an agent reply and exit once the server accepts it, when you are not about to long-poll - `npx -y lavish-axi design` for design-direction priority and current snippets - `npx -y lavish-axi playbook <id>` for focused artifact guidance (`npx -y lavish-axi playbook` lists ids) You do not need lavish-axi installed globally - invoke it with `npx -y lavish-axi <html-file>`. If lavish-axi output shows a follow-up command starting with `lavish-axi`, run it as `npx -y lavish-axi ...` instead. ## Request $ARGUMENTS If the request above is non-empty, the user invoked `/lavish` explicitly - fetch the current CLI guidance, then build that artifact as an HTML file. If it is empty, infer what to visualize from the conversation.