← ClaudeAtlas

panerelay-browserlisted

Use agent-browser through Panerelay to work in the user's existing Chrome or Microsoft Edge session and explicitly authorized tabs. Use when the user asks to operate their current browser, reuse an existing login or cookies, work through Panerelay or its side panel, or control a tab they authorized.
F-loat/panerelay · ★ 31 · AI & Automation · score 80
Install: claude install-skill F-loat/panerelay
# Panerelay Browser Use standard `agent-browser` commands with the `panerelay` browser Provider. Panerelay supplies the connection to the user's browser; agent-browser retains browser automation semantics. Chrome is the verified runtime baseline. Microsoft Edge operations use the shared Chromium implementation and remain `Forwarded` pending complete representative acceptance; see the [browser platform compatibility record](https://github.com/F-loat/panerelay/blob/main/docs/compatibility/browser-platforms.md). ## Workflow 1. Before the first browser command, load the installed agent-browser core skill when available: ```bash agent-browser skills get core ``` 2. Use one stable session and pass the Provider explicitly: ```bash agent-browser --session panerelay-task --provider panerelay snapshot -i agent-browser --session panerelay-task --provider panerelay click @e1 ``` A user or project configured with `npx --yes @panerelay/setup --global-provider` or `npx --yes @panerelay/setup --project-provider` already defaults to this Provider, but keep the explicit flag when portability matters. Explicit CLI selection wins over project and user defaults. Changing the Provider never changes browser authorization. 3. If Panerelay reports multiple ready browsers, inspect them with `npx --yes @panerelay/cli browsers`. Ask the user which browser to use when their intent is not already explicit, then scope the process with `PANERELAY_BROWSER_ID=<registration-id>