Provides a universal web dashboard for OpenClaw instances: schema-driven UIs, forms, tables, charts and an AI chat panel to manage and interact with installed skills. Includes user management, role-based access control, audit logging, HTTPS setup with Let's Encrypt, and maintenance tasks.
Use Webclaw when you need a browser-accessible admin interface to view and operate installed skills, manage users and roles, enforce HTTPS, or run site-wide maintenance (session cleanup, cert renewal). Ideal for non-CLI users and teams that prefer a GUI.
Best suited for agents with system/infra capabilities (agents that can run Bash, manage services and edit configs), e.g., Codex/Copilot-style agents or On-host automation agents.
WebClaw is a web extraction skill with 12 API endpoints for scraping, crawling, extracting, summarizing, searching, and monitoring web content with automatic antibot bypass. No bundled scripts — it operates purely via API calls using WEBCLAW_API_KEY. The SKILL.md is comprehensive with clear endpoint documentation, request/response examples, and usage tips. API key is passed via environment variable (good practice), and all network calls go to a single known API host (api.webclaw.io). No destructive commands, no exfiltration instructions, no shell injection risks.
Well-documented API skill with no scripts. All traffic goes to api.webclaw.io via HTTPS. No hardcoded credentials, no destructive commands, no shell injection vectors. The skill is essentially a well-structured API reference. Minor deduction for telemetry aspect: the skill directs all web scraping through a third-party service which could log URLs scraped. Architecture is solid but monolithic — one large SKILL.md with all 12 endpoints could benefit from splitting into references/. Code quality is good with clear instructions, good examples, and helpful comparison tables, but lacks error handling guidance or troubleshooting section.