Browser automation for DeepSeek Harness via bsk — lets your agent navigate, snapshot, click, fill, and screenshot the real Chromium browser (with your logins).
dsh plugin add dsh-browserRequires bsk on PATH + the browser-skill extension loaded (see browser-skill prerequisites).
| Tool | What it does |
|---|---|
browse_navigate |
Open a URL |
browse_snapshot |
Read the page as an accessibility tree (with @eN refs) |
browse_click |
Click an element by @eN ref or CSS selector |
browse_fill |
Type into an input |
browse_screenshot |
Capture a screenshot |
browse_observe |
Semantic VOM observation |
Open https://example.com, read the page, and tell me the title.
MIT