Websites · PROJECT NOTES
BrowserART
A browser-agent safety evaluation toolkit with synthetic websites, behavior datasets and trajectory classifiers for studying whether agents follow harmful requests.
What can you do with it?
Evaluate browser-agent refusal behavior in a controlled research setup and inspect recorded trajectories.
What should you know before using it?
A specialized safety benchmark rather than a general task-success ranking. It mixes synthetic and live-site tasks; the code is MIT-licensed while the behavior dataset lists CC BY-NC 4.0.
What is verified here?
This entry was reviewed against original project material. We have not independently run the software or reproduced its results. AI-development claims are author-reported; check upstream for current versions, setup requirements, and licenses.
- Resource type
- Environment + benchmark
- Provenance
- Development provenance not established
- Reviewed
- Platforms
- Web