Skip to content
Agentmart
Submit a tool

agent-browser

Use when you need to reproduce or debug web UI flows (especially auth/OIDC) via the Agent Browser CLI, capture snapshots/screenshots, and extract redirect…

Skill by intelligentcode-ai

IndexedMIT★ 1
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/intelligentcode-ai-agent-browser

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Use when you need to reproduce or debug web UI flows (especially auth/OIDC) via the Agent Browser CLI, capture snapshots/screenshots, and extract redirect URLs and on-page errors deterministically. Includes install/setup guidance when the CLI is missing.

Use cases: design, security, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/agent-browser

What's inside

Skill name
agent-browser
Description
Use when you need to reproduce or debug web UI flows (especially auth/OIDC) via the Agent Browser CLI, capture snapshots/screenshots, and extract redirect URLs and on-page errors deterministically. Includes install/setup guidance when the CLI is missing.
License
MIT
Instructions
178 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# Agent Browser (CLI)

Use this skill when we need to verify UI behavior ourselves (login flows, redirects, error toasts) without guessing.

## Install / Setup (If `agent-browser` Is Missing)

1. Check if the CLI is available:

```bash
command -v agent-browser && agent-browser --help | head
```

2. Install Agent Browser CLI (pick one):

macOS (Homebrew):

```bash
brew install agent-browser
```

Linux:
- First install Node.js + npm (examples):

```bash
# Debian/Ubuntu
sudo apt update && sudo apt install -y nodejs npm

# Fedora
sudo dnf install -y nodejs npm

# Arch
sudo pacman -S --needed nodejs npm
```

If your distro Node.js is too old, use a version manager (nvm/fnm/asdf) and install a current Node LTS.

All platforms (npm, recommended/fallback):

```bash
npm install -g agent-browser
```

Windows note:
- If you don't have Node.js yet, install it first (PowerShell):

```powershell
winget install OpenJS.NodeJS.LTS
```

3. Install the browser runtime the CLI uses (one-time per machine/user):

```bash
agent-browser install
```

Linux-only note:
- If the browser fails to launch due to missing system libraries, try:

```bash
agent-browser install --with-deps
```

4. Sanity check:

```bash
agent-browser open https://example.com/
agent-browser snapshot -i
agent-browser close
```

## Artifacts (Cross-Platform)

Prefer saving evidence inside the repo so it can be attached/shared predictably:

- Default: `.agent/artifacts/agent-browser/<run-id>/`
- Include: screenshots (`.png`) and snapshots (`.txt`)

Example (macOS/Linux):

```bash
run_id="oidc-$(date +%Y%m%d-%H%M%S)"
out_dir=".agent/artifacts/agent-browser/$run_id"
mkdir -p "$out_dir"
echo "$out_dir"
```

Example (Windows PowerShell):

```powershell
$runId = "oidc-" + (Get-Date -Format "yyyyMMdd-HHmmss")
$outDir = ".agent/artifacts/agent-browser/$runId"
New-Item -ItemType Directory -Force -Path $outDir | Out-Null
$outDir
```

## Quick Start (OIDC Debug)

The core loop is: `open` -> `snapshot -i` -> interact (prefer refs like `@e2`) -> re-`snapshot -i` -> capture evidence.

### macOS/Linux (bash)

1. Start a clean session, open the app, and capture baseline evidence:

```bash
run_id="oidc-$(date +%Y%m%d-%H%M%S)"
out_dir=".agent/artifacts/agent-browser/$run_id"
mkdir -p "$out_dir"

agent-browser close || true
agent-browser --session oidc open https://example.com/   # replace
agent-browser --session oidc wait 1000
agent-browser --session oidc screenshot "$out_dir/01-home.png"
agent-browser --session oidc snapshot -i > "$out_dir/01-home.snapshot.txt"
```

2. Prefer clicking by ref from the snapshot (fallback to semantic find if needed):

```bash
# Look in 01-home.snapshot.txt for: button "Sign in" [ref=e2]

From intelligentcode-ai/skills, MIT.

More from intelligentcode-ai/skills

Signals

Stars
1
Forks
1
Open issues
0
Last commit (this folder)
Feb 14, 2026 (8 months ago)
Last release
Feb 16, 2026 (8 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from intelligentcode-ai

Cite this page

agent-browser skill. Agent Mart. https://agentmart.dev/skills/intelligentcode-ai-agent-browser (last updated 2026-10-24)