Skills
How an AI agent uses ego (lite)'s browser tools.
The ego-browser skill shows your AI agent how to use ego (lite). With it, the agent can open pages in a Space, read Snapshots, and use ego-browser tools to complete a task.
Use the skill
Install ego (lite) and complete setup. Open your agent and check whether ego-browser appears in its skills. Select it, then describe what you want the agent to do. If the skill is missing, close and reopen the agent after setup before checking again. Review any request to run ego-browser or open ego (lite) before continuing.
The Quick start has a task you can try.
For the skill picker and permission settings in each Windows agent, follow its tutorial:
- Claude Code
- Codex
- OpenCode: select
@ego-browserin the input box. - WorkBuddy: select the skill, then enter the task in the same message.
- Hermes Agent: enter
/ego-browser, add a space, and write the task in the same message. - OpenClaw: enter
/skill ego-browser, add a space, and write the task in the same message. - DeepSeek Harness: select
/ego-browser, then enter the task in the same message. - Your own Agent Harness: check local command access and load the skill through your Harness's supported interface.
- Local model: configure Qwen with Ollama, connect your Harness, and load ego-browser.
Stay in control
The skill lets an agent interact with real web pages. Watch for permission prompts, and tell the agent when to pause for your input. You should handle sign-ins, verification codes, payments, and final submission decisions yourself.