Skill Detail

Run local multi-agent browser tasks with Nanobrowser

Use Nanobrowser when an operator needs a local Chrome or Edge browser agent to run multi-step web tasks with their own LLM keys and visible browser state.

Browser AutomationMulti-Framework
Browser Automation Multi-Framework Security Reviewed
⭐ 13.1k GitHub stars
INSTALL WITH ANY AGENT
npx skills add agentskillexchange/skills --skill run-local-multi-agent-browser-tasks-with-nanobrowser Copy
Works best when you want a reusable capability, not another fragile one-off prompt.
At a glance
Tools required
Chrome or Edge, Nanobrowser extension, user-managed LLM API key, supported model provider
Install & setup
Install Nanobrowser from the Chrome Web Store or download the release zip, load it into Chrome or Edge, open the extension settings, add LLM API keys, and assign models to the planner and navigator agents.
Author
Nanobrowser
Publisher
Open Source Project
Last updated
Jun 2, 2026
Quick brief

Use Nanobrowser for browser tasks where the agent must operate inside the user’s own Chrome or Edge session, keep credentials local, and expose the navigation process for review. The operator installs the extension, configures provider keys, assigns models to the planner and navigator agents, then gives a concrete web task such as collecting information, completing a repetitive workflow, or inspecting a site. During the run, the operator can monitor browser state, review failures, adjust the task prompt, and rerun the job with narrower instructions. This is not a generic browser extension listing: the skill boundary is the repeatable operator workflow of configuring local browser agents, running a scoped task, watching the intermediate browser actions, and deciding whether the output is acceptable before using it downstream. Invoke it instead of using the product normally when the value is delegated, auditable web operation rather than manual browsing. Keep the task limited to sites and accounts the user is authorized to operate.