Back to Skills

opencli-explorer

How to automate Antigravity using OpenCLI

18,693stars1,835forksUpdated 5/5/2026

Security Assessment

Low Risk(70/100)

Detected risks:

Secret Exposure([references/adapter-templates.md] token =, [references/advanced-patterns.md] token =)
Security Score70/100

About opencli-explorer

The opencli-explorer skill is designed to guide users through the complete process of building an OpenCLI adapter from scratch by systematically discovering and integrating web APIs. It addresses the common challenge of extracting structured data from modern websites, especially those that rely on dynamic loading, authentication, or hidden endpoints. Instead of relying on static inspection, this skill enforces a browser-driven exploration workflow to uncover real API behavior, ensuring that adapters are based on accurate, reproducible network interactions.

Its core capabilities include a structured four-step workflow: API discovery, authentication strategy selection, adapter implementation, and testing. The skill provides a browser-based exploration method using DevTools-like commands to inspect elements, capture network requests, simulate user interactions, and validate API responses. It also outlines multiple API discovery techniques such as JSON endpoint probing, global state inspection, and interaction-triggered loading. Additionally, it helps users determine the appropriate authentication approach, including public access, cookie-based sessions, or token-based headers, ensuring compatibility with different platform security models.

This skill is particularly useful for developers building CLI tools for websites, automation engineers working on web scraping or API integration, and contributors developing adapters for OpenCLI ecosystems. It supports both one-off CLI generation and scalable multi-command adapter development. Whether targeting a new platform, reverse-engineering undocumented APIs, or preparing a production-ready adapter for contribution, opencli-explorer provides a practical and methodical framework for achieving reliable results.

FAQ

Do I need to use a browser to discover APIs?

Yes. The workflow explicitly requires browser-based exploration to capture real network requests, especially for APIs triggered by user interactions or lazy loading.

Can this skill work with any website?

It works with most modern websites, but effectiveness depends on whether APIs can be observed via browser network activity and whether authentication can be reproduced.

What if the API returns empty data despite a successful response?

This may indicate missing authentication such as cookies or tokens. The skill recommends verifying headers and checking for anti-bot or signature requirements.

Is this suitable for generating a full CLI automatically?

It supports automation attempts via commands like generation tools, but manual exploration is recommended if automation fails or for complex sites.

What kind of authentication methods are supported?

The skill covers public access, cookie-based authentication, and token-based approaches such as Bearer tokens stored in localStorage.

All Files

4 files
references/record-workflow.md3.2 KB
View
references/adapter-templates.md12.3 KB
View
references/advanced-patterns.md6.6 KB
View
SKILL.md8.5 KB
View

Install opencli-explorer

Download and extract the skill files to your .claude/skills/ directory.

Quick Setup:

  1. Copy the skill folder to .claude/skills/
  2. Claude will automatically detect and use the skill