opencli-web-automation
Turn any website into a CLI using browser session reuse and AI-powered command discovery
Best use case
opencli-web-automation is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Turn any website into a CLI using browser session reuse and AI-powered command discovery
Teams using opencli-web-automation should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/opencli-web-automation/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How opencli-web-automation Compares
| Feature / Agent | opencli-web-automation | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Turn any website into a CLI using browser session reuse and AI-powered command discovery
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# OpenCLI Web Automation
> Skill by [ara.so](https://ara.so) — Daily 2026 Skills collection.
OpenCLI turns any website into a command-line interface by reusing Chrome's logged-in browser session. It supports 19 sites and 80+ commands out of the box, and lets you add new adapters via TypeScript or YAML dropped into the `clis/` folder.
---
## Installation
```bash
# Install globally via npm
npm install -g @jackwener/opencli
# One-time setup: discovers Playwright MCP token and distributes to all tools
opencli setup
# Verify everything is working
opencli doctor --live
```
### Prerequisites
- Node.js >= 18.0.0
- Chrome browser **running and logged into the target site**
- [Playwright MCP Bridge](https://chromewebstore.google.com/detail/playwright-mcp-bridge/mmlmfjhmonkocbjadbfplnigmagldckm) extension installed in Chrome
### Install from Source (Development)
```bash
git clone git@github.com:jackwener/opencli.git
cd opencli
npm install
npm run build
npm link
```
---
## Environment Configuration
```bash
# Required: set in ~/.zshrc or ~/.bashrc after running opencli setup
export PLAYWRIGHT_MCP_EXTENSION_TOKEN="<your-token-from-setup>"
```
MCP client config (Claude/Cursor/Codex `~/.config/*/config.json`):
```json
{
"mcpServers": {
"playwright": {
"command": "npx",
"args": ["-y", "@playwright/mcp@latest", "--extension"],
"env": {
"PLAYWRIGHT_MCP_EXTENSION_TOKEN": "$PLAYWRIGHT_MCP_EXTENSION_TOKEN"
}
}
}
}
```
---
## Key CLI Commands
### Discovery & Registry
```bash
opencli list # Show all registered commands
opencli list -f yaml # Output registry as YAML
opencli list -f json # Output registry as JSON
```
### Running Built-in Commands
```bash
# Public API commands (no browser login needed)
opencli hackernews top --limit 10
opencli github search "playwright automation"
opencli bbc news
# Browser commands (must be logged into site in Chrome)
opencli bilibili hot --limit 5
opencli twitter trending
opencli zhihu hot -f json
opencli reddit frontpage --limit 20
opencli xiaohongshu search "TypeScript"
opencli youtube search "browser automation"
opencli linkedin search "senior engineer"
```
### Output Formats
All commands support `--format` / `-f`:
```bash
opencli bilibili hot -f table # Rich terminal table (default)
opencli bilibili hot -f json # JSON (pipe to jq)
opencli bilibili hot -f yaml # YAML
opencli bilibili hot -f md # Markdown
opencli bilibili hot -f csv # CSV export
opencli bilibili hot -v # Verbose: show pipeline debug steps
```
### AI Agent Workflow (Creating New Commands)
```bash
# 1. Deep explore a site — discovers APIs, auth, capabilities
opencli explore https://example.com --site mysite
# 2. Synthesize YAML adapters from explore artifacts
opencli synthesize mysite
# 3. One-shot: explore → synthesize → register in one command
opencli generate https://example.com --goal "hot posts"
# 4. Strategy cascade — auto-probes PUBLIC → COOKIE → HEADER auth
opencli cascade https://api.example.com/data
```
Explore artifacts are saved to `.opencli/explore/<site>/`:
- `manifest.json` — site metadata
- `endpoints.json` — discovered API endpoints
- `capabilities.json` — inferred command capabilities
- `auth.json` — authentication strategy
---
## Adding a New Adapter
### Option 1: YAML Declarative Adapter
Drop a `.yaml` file into `clis/` — auto-registered on next run:
```yaml
# clis/producthunt.yaml
site: producthunt
commands:
- name: trending
description: Get trending products on Product Hunt
args:
- name: limit
type: number
default: 10
pipeline:
- type: navigate
url: https://www.producthunt.com
- type: waitFor
selector: "[data-test='post-item']"
- type: extract
selector: "[data-test='post-item']"
fields:
name:
selector: "h3"
type: text
tagline:
selector: "p"
type: text
votes:
selector: "[data-test='vote-button']"
type: text
url:
selector: "a"
attr: href
- type: limit
count: "{{limit}}"
```
### Option 2: TypeScript Adapter
```typescript
// clis/producthunt.ts
import type { CLIAdapter } from "../src/types";
const adapter: CLIAdapter = {
site: "producthunt",
commands: [
{
name: "trending",
description: "Get trending products on Product Hunt",
options: [
{
flags: "--limit <n>",
description: "Number of results",
defaultValue: "10",
},
],
async run(options, browser) {
const page = await browser.currentPage();
await page.goto("https://www.producthunt.com");
await page.waitForSelector("[data-test='post-item']");
const products = await page.evaluate(() => {
return Array.from(
document.querySelectorAll("[data-test='post-item']")
).map((el) => ({
name: el.querySelector("h3")?.textContent?.trim() ?? "",
tagline: el.querySelector("p")?.textContent?.trim() ?? "",
votes:
el
.querySelector("[data-test='vote-button']")
?.textContent?.trim() ?? "",
url:
(el.querySelector("a") as HTMLAnchorElement)?.href ?? "",
}));
});
return products.slice(0, Number(options.limit));
},
},
],
};
export default adapter;
```
---
## Common Patterns
### Pattern: Authenticated API Extraction (Cookie Injection)
```typescript
// When a site exposes a JSON API but requires login cookies
async run(options, browser) {
const page = await browser.currentPage();
// Navigate first to ensure cookies are active
await page.goto("https://api.example.com");
const data = await page.evaluate(async () => {
const res = await fetch("/api/v1/feed?limit=20", {
credentials: "include", // reuse browser cookies
});
return res.json();
});
return data.items;
}
```
### Pattern: Header Token Extraction
```typescript
// Extract auth tokens from browser storage for API calls
async run(options, browser) {
const page = await browser.currentPage();
await page.goto("https://example.com");
const token = await page.evaluate(() => {
return localStorage.getItem("auth_token") ||
sessionStorage.getItem("token");
});
const data = await page.evaluate(async (tok) => {
const res = await fetch("/api/data", {
headers: { Authorization: `Bearer ${tok}` },
});
return res.json();
}, token);
return data;
}
```
### Pattern: DOM Scraping with Wait
```typescript
async run(options, browser) {
const page = await browser.currentPage();
await page.goto("https://news.ycombinator.com");
// Wait for dynamic content to load
await page.waitForSelector(".athing", { timeout: 10000 });
return page.evaluate((limit) => {
return Array.from(document.querySelectorAll(".athing"))
.slice(0, limit)
.map((row) => ({
title: row.querySelector(".titleline a")?.textContent?.trim(),
url: (row.querySelector(".titleline a") as HTMLAnchorElement)?.href,
score:
row.nextElementSibling
?.querySelector(".score")
?.textContent?.trim() ?? "0",
}));
}, Number(options.limit));
}
```
### Pattern: Pagination
```typescript
async run(options, browser) {
const page = await browser.currentPage();
const results = [];
let pageNum = 1;
while (results.length < Number(options.limit)) {
await page.goto(`https://example.com/posts?page=${pageNum}`);
await page.waitForSelector(".post-item");
const items = await page.evaluate(() =>
Array.from(document.querySelectorAll(".post-item")).map((el) => ({
title: el.querySelector("h2")?.textContent?.trim(),
url: (el.querySelector("a") as HTMLAnchorElement)?.href,
}))
);
if (items.length === 0) break;
results.push(...items);
pageNum++;
}
return results.slice(0, Number(options.limit));
}
```
---
## Maintenance Commands
```bash
# Diagnose token and config across all tools
opencli doctor
# Test live browser connectivity
opencli doctor --live
# Fix mismatched configs interactively
opencli doctor --fix
# Fix all configs non-interactively
opencli doctor --fix -y
```
---
## Testing
```bash
npm run build
# Run all tests
npx vitest run
# Unit tests only
npx vitest run src/
# E2E tests only
npx vitest run tests/e2e/
# Headless browser mode for CI
OPENCLI_HEADLESS=1 npx vitest run tests/e2e/
```
---
## Troubleshooting
| Symptom | Fix |
|---|---|
| `Failed to connect to Playwright MCP Bridge` | Ensure extension is enabled in Chrome; restart Chrome after install |
| Empty data / `Unauthorized` | Open Chrome, navigate to the site, log in or refresh the page |
| Node API errors | Upgrade to Node.js >= 18 |
| Token not found | Run `opencli setup` or `opencli doctor --fix` |
| Stale login session | Visit the target site in Chrome and interact with it to prove human presence |
### Debug Verbose Mode
```bash
# See full pipeline execution steps
opencli bilibili hot -v
# Check what explore discovered
cat .opencli/explore/mysite/endpoints.json
cat .opencli/explore/mysite/auth.json
```
---
## Project Structure (for Adapter Authors)
```
opencli/
├── clis/ # Drop .ts or .yaml adapters here (auto-registered)
│ ├── bilibili.ts
│ ├── twitter.ts
│ └── hackernews.yaml
├── src/
│ ├── types.ts # CLIAdapter, Command interfaces
│ ├── browser.ts # Playwright MCP bridge wrapper
│ ├── loader.ts # Dynamic adapter loader
│ └── output.ts # table/json/yaml/md/csv formatters
├── tests/
│ └── e2e/ # E2E tests per site
└── CLI-EXPLORER.md # Full AI agent exploration workflow
```Related Skills
opencli-rs-web-scraper
Blazing fast Rust CLI tool to fetch information from 55+ websites (Twitter, Reddit, YouTube, Bilibili, Zhihu, etc.) with a single command, plus AI-native discovery and Electron app control.
maante-game-automation
MaaNTE is a MAA-based automation assistant for the game Neverness to Everness, powered by MaaFramework, supporting auto-fishing, auto-coffee-making, and cafe revenue extraction.
geoflow-content-automation
GEOFlow open-source GEO/SEO content production system with AI generation, review workflow, and publishing pipeline built on PHP and PostgreSQL.
codex-oauth-automation-extension
Chrome extension for automating OpenAI OAuth registration flows with captcha retrieval, CPA callback verification, and auto-recovery across multiple rounds
agent-browser-automation
Headless browser automation CLI for AI agents using native Rust binary with Chrome DevTools Protocol
```markdown
---
zeroboot-vm-sandbox
Sub-millisecond VM sandboxes for AI agents using copy-on-write KVM forking via Zeroboot
yourvpndead-vpn-detection
Android app that detects VPN/proxy servers (VLESS/xray/sing-box) via local SOCKS5 vulnerability, exposing exit IPs and server configs without root
xata-postgres-platform
Expert skill for Xata open-source cloud-native Postgres platform with copy-on-write branching, scale-to-zero, and Kubernetes deployment
x-mentor-skill-nuwa
AI-powered X (Twitter) content strategy skill that distills methodologies from 6 top creators + open-source algorithm data into actionable writing, growth, and monetization guidance.
wx-favorites-report
End-to-end pipeline to extract, decrypt, and visualize WeChat Mac favorites from encrypted SQLite DB into an interactive HTML report.
wterm-web-terminal
Web terminal emulator with Zig/WASM core, DOM rendering, and React/vanilla JS bindings