galiprandi/job-seeker / .agents/skills/browser-automation
browser-automation
VerifiedControl a dedicated browser via agent-browser (Vercel). Use when automating web apps, scraping authenticated sites, filling forms, or navigating SPAs. Do NOT use for desktop apps or API-only integrations.
SETUP.md
View on GitHub# Setup — browser-automation (agent-browser engine)
Requirements to run this skill on a new machine.
## Requirements
- **Node.js** 18+
- **System Google Chrome** (`google-chrome`, `chromium`, or `chromium-browser`). Required: agent-browser's bundled Chromium fails UA-gates on hostile sites (WhatsApp, Google sign-in).
- **agent-browser** (Vercel CLI) — optional: the wrapper falls back to `npx -y agent-browser` automatically. Global install only saves the per-call npx lookup:
```bash
npm install -g agent-browser
agent-browser --version # verify (or let the wrapper use npx)
```
Without a global install, `npx agent-browser <cmd>` works too.
## Configuration
- `agent-browser.json` at the consuming repo root: committed, repo-scoped defaults (namespace, profile, headed, downloads, screenshots). The skill ships a template (`scripts/agent-browser.template.json`); the wrapper auto-copies it into the repo root on first use, substituting the repo name as namespace — never overwrites an existing file.
- Env vars the wrapper sets on its own (do not configure by hand):
- `AGENT_BROWSER_PROFILE` → repo's `.browser-profile/`
- `AGENT_BROWSER_NAMESPACE` → repo directory name (per-repo daemon isolation)
- `AGENT_BROWSER_ARGS` → `--no-sandbox,--disable-blink-features=AutomationControlled`
- `AGENT_BROWSER_EXECUTABLE_PATH` → autodetects system Chrome
- `AGENT_BROWSER_CONFIG` → repo's `agent-browser.json`
## First run
```bash
node .agents/skills/browser-automation/scripts/browser.js open https://example.com --headed
```
- Sessions (logins) persist in `.browser-profile/` — same directory as the previous engine; the profile is 100% compatible.
- **Never run two browsers on the same profile at once** (SingletonLock).
- **Google sites (Gmail, Calendar, Drive):** the login must happen INSIDE the automated browser (headed) — Google rejects sign-ins inherited from regular Chrome in a CDP context. Once logged in, the session persists.
- **All authentication is manual login inside the dedicated browser.** No passwords through chat channels, no credential vaults — the Chrome profile is the only credential store.
## Diagnostics — `agent-browser doctor`
Runs the full install check:
```bash
agent-browser doctor # full diagnostics
agent-browser doctor --quick # fast, no launch test
agent-browser doctor --fix # repairs detected issues
```
**Expected failures on this setup (not real problems):**
- `Chrome for Testing CDN unreachable` → we don't use bundled Chrome (UA-gates), we use system Chrome. Ignore.
- `Browser launch failed: No usable sandbox` → doctor's launch test runs without `--no-sandbox`. The wrapper does pass it, so real usage works. Ignore unless real usage also fails.
**Expected result on this machine:** 10 pass, 2 fail (the two above).
- Chrome won't launch on Linux/AppArmor → `--no-sandbox` (already in the wrapper).
- WhatsApp or other UA-gates fail → verify `AGENT_BROWSER_EXECUTABLE_PATH` points to real Chrome, not the bundled one.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |