Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.
90
88%
Does it follow best practices?
Impact
92%
2.62xAverage score across 6 eval scenarios
Critical
Do not install without reviewing
The canonical home for this skill is agent-browser in vercel-labs/agent-browser
Auth vault and secure credential management
auth save with --password-stdin
0%
100%
auth login used
0%
100%
Content boundaries enabled
0%
100%
Encryption key used
0%
100%
Named session-name used
0%
100%
No inline password
100%
100%
State files not committed
0%
0%
Content boundary explanation
100%
100%
get text used for extraction
0%
100%
Browser closed
0%
100%
Parallel sessions and state persistence
Named sessions used
0%
100%
Unique session per site
0%
100%
Concurrent launch
100%
100%
state save used
0%
100%
Three state files created
100%
100%
MAX_OUTPUT set
0%
100%
All sessions closed
0%
100%
snapshot -i used per session
0%
0%
Content files created
100%
100%
session_report.md present
100%
100%
File downloads and semantic locators
Semantic locators used
100%
100%
find text or find role present
100%
100%
Download command used
100%
100%
Download path configured
100%
100%
agent-browser.json created
100%
100%
MAX_OUTPUT set
100%
100%
wait --load networkidle used
100%
100%
Browser closed at end
100%
100%
snapshot -i used
0%
100%
fetch_log.txt produced
100%
100%
Visual regression and page diffing
diff url used
0%
100%
Two pages compared
0%
100%
diff screenshot --baseline used
0%
100%
Dark mode screenshot taken
50%
100%
Light and dark screenshots both present
100%
100%
record start used
0%
100%
record stop used
0%
100%
wait --load networkidle used
0%
100%
Browser closed at end
0%
100%
qa_report.md contains diff output references
100%
100%
Core browser workflow and verification
Uses snapshot -i
0%
100%
Wait networkidle used
0%
100%
Command chaining with &&
0%
0%
Re-snapshot after navigation
0%
0%
diff snapshot used
0%
100%
diff output saved
40%
100%
Browser closed
0%
100%
City text files created
100%
100%
get text used for extraction
0%
100%
Process log present
60%
100%
JavaScript evaluation and visual element interaction
eval --stdin or -b used
0%
100%
Complex JS not inline-quoted
100%
100%
Annotated screenshot taken
0%
100%
Scoped snapshot used
0%
100%
Domain allowlist set
0%
100%
snapshot -i used
0%
0%
wait --load networkidle used
0%
100%
audit_results.json exists
100%
100%
annotated.png exists
100%
100%
main_snapshot.txt exists
100%
100%
0e57985
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.