Anthropic Brings Claude Cowork to Its Chrome Extension, Adding Skills and Plugins to the Browser
Anthropic has launched Claude Cowork, a new feature for its Chrome extension that turns the AI assistant into an active browser collaborator. Instead of simply reading webpages, Claude can now execute tasks, fill forms, and interact with third-party plugins directly within the browser.
This update transforms Claude from a passive reading tool into an agentic assistant that works alongside users. The feature is designed to help with repetitive online tasks while maintaining Claude’s safety guardrails.
What Claude Cowork Does
Claude Cowork builds on the existing Chrome extension’s ability to read and summarize webpage content. The key shift: Claude can now click buttons, type text, and navigate multi-step workflows on your behalf.
- Fills forms and submits data automatically, including job applications, sign-ups, or data entry.
- Navigates complex websites by clicking through menus, tabs, and links to find specific information.
- Interacts with third-party plugins such as password managers, form fillers, or developer tools.
- Executes multi-step tasks like booking appointments or ordering supplies without manual intervention.
The feature works across any webpage, not just sites Anthropic has pre-configured.
How Plugins and Skills Work
Anthropic introduces a plugin system for Claude Cowork, allowing external tools to register actions Claude can invoke.
- Skills are modular capabilities that Claude can learn, like “fill this loan application” or “check today’s stock prices.”
- Plugins extend what Claude can do by providing APIs or UI actions the assistant can trigger.
- Developers can build custom skills using a new Anthropic SDK, enabling enterprise workflows.
- Third-party plugin partners include productivity, financial, and research tool providers.
Users can enable or disable specific skills and plugins in the Chrome extension settings, giving control over which capabilities are active.
Safety and Privacy Considerations
Anthropic emphasizes that Claude Cowork operates within the same safety framework as its chat and API products. The assistant cannot bypass user permissions or act without confirmation.
- All actions require user approval before execution, preventing unauthorized changes.
- Claude asks for confirmation on high-risk actions like deleting data or sending emails.
- Sensitive fields are protected by default, requiring explicit user permission to fill credit card numbers or passwords.
- Session logging lets users review what Claude did after each task.
The company also notes that plugins are sandboxed and permission-gated, preventing one plugin from accessing data from another without consent.
Comparison to Competitors
Claude Cowork enters a crowded field of browser AI assistants. Google’s Gemini offers similar in-browser task automation, and OpenAI’s ChatGPT has a plugin ecosystem for web interactions.
- Anthropic’s advantage is safety-first design with explicit user confirmations on every action.
- Gemini is deeper integrated into Google’s ecosystem but has broader data access.
- ChatGPT plugins are more mature but have faced scrutiny over data handling.
- Claude Cowork is currently Chrome-only, with no word on Firefox or Edge support.
Anthropic positions this as a trustworthy alternative for users who want automation without sacrificing control.
Availability and Rollout
Claude Cowork is available today to all Claude Pro and Claude Team subscribers using the Chrome extension. Free-tier users can try a limited version with fewer skills.
- Rollout is gradual over the next week to ensure stability.
- European users get access immediately, with GDPR compliance confirmed.
- Enterprise customers can request custom skill development through Anthropic’s sales team.
- No additional cost beyond the existing subscription fee.
Anthropic plans to expand the plugin marketplace and add more skills based on user feedback.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.