VulpineOS
Docs
Runtime
Runtime overview
Hardened browser, orchestrator, MCP
Injection filter
Strip invisible DOM before the agent
Action lock
Freeze JS, timers, reflows mid-think
Optimised DOM
93.1% measured token reduction
Operator surfaces
Web panel
Embedded operator console
TUI workbench
SSH-friendly terminal UI
MCP toolbelt
36 typed browser tools
Vulpine-Box
One-container self-host
Ecosystem
Foxbridge
CDP ↔ Juggler / WebDriver BiDi
vulpine-mark
Set-of-Mark visual labelling
mobilebridge
Android device sessions
OpenClaw
Pre-configured agent loop
Documentation
Quickstart
First agent in 30 seconds
Architecture
Four-phase security model
MCP Reference
All 36 browser tools
Agent Scripting DSL
Declarative agent definitions
Open source
VulpineOS Runtime
GitHub, MPL 2.0
Foxbridge
CDP ↔ Juggler / WebDriver BiDi
vulpine-mark
Set-of-Mark visual labelling
mobilebridge
Android device discovery + sessions
Integrations
OpenClaw
Pre-configured agent loop
Camoufox
Anti-detect browser core
LLM providers
30+ models, one toolbelt
Docker (Vulpine-Box)
One-click self-host
Guides
AI Agent Security
Threat model, mitigations
Prevent Prompt Injection
Hidden DOM, ARIA tricks
OpenClaw + Camoufox
End-to-end setup
Camoufox vs Chrome
Why Firefox, why patches
Engineering
Changelog
Release notes as work lands
Roadmap
Now, next, later, research
Blog
Long-form posts on the runtime
RenderLab
Optional render-diagnostics study
Community
Support
How to get help, fast
GitHub Discussions
Ideas, RFCs, support
Contributing
How to help
Security
Responsible disclosure
Star on GitHub →
Security · Last updated 9 May 2026

Responsible disclosure.

How to report a security issue in VulpineOS, what we treat as in scope, the safe-harbor language we operate under, and the response windows you can expect.

01

Reporting a security issue

Email security@vulpineos.com. For high-impact issues, use the subject line VULN: short title so we can triage faster. We will acknowledge within 2 business days and follow up with the next steps.

For lower-severity, code-level issues you can also open a private security advisory on the relevant GitHub repository:

  • VulpineOS runtime
  • Foxbridge
02

Scope

In scope:

  • vulpineos.com (marketing site)
  • auth.vulpineos.com (waitlist + login)
  • admin.vulpineos.com (operator console)
  • The VulpineOS, Foxbridge, vulpine-mark, and mobilebridge open-source repositories
  • The Vulpine RenderLab consent + diagnostics flow
03

Out of scope

  • Third-party services we use (Vercel, Supabase, Google) — please report those to the upstream provider directly.
  • Self-hosted instances of the open-source runtime that you operate on your own infrastructure.
  • Reports based purely on best-practice headers (HSTS preload, CSP nonces, SPF / DKIM strictness) without a demonstrable security impact.
  • Volumetric denial-of-service, social engineering of staff, and physical access attempts.
  • Vulnerabilities in versions that are no longer the latest published release.
04

Safe harbor

If you make a good-faith effort to comply with this policy, do not access more data than necessary to demonstrate the issue, do not exfiltrate data, do not publicly disclose the issue before we have had a reasonable chance to fix it, and do not disrupt other users, we will:

  • Not pursue legal action against you for that research.
  • Work with you on disclosure timing and credit.
  • Treat your report as a private security advisory until a fix is shipped and users have had time to update.
05

Response expectations

  • Acknowledgement within 2 business days.
  • Initial assessment within 5 business days.
  • Patch + advisory window targeted at 30 days for high-severity issues, faster where the impact justifies it.
  • Public credit on the advisory if you want it (we are happy to coordinate attribution and CVE assignment via the GitHub Security Advisory flow).
06

Encrypted reporting

For sensitive reports, request a PGP key by emailing security@vulpineos.com with the subject PGP REQUEST. We will share the current public key.

07

Abuse, not vulnerabilities

Reports of abuse — someone using VulpineOS to violate our Acceptable Use Policy — should go to abuse@vulpineos.com instead.

See also
  • Terms of Service
  • Privacy Policy
  • Cookies
  • Acceptable use
  • Subprocessors
VulpineOS

The browser built for AI agents.
Open-source runtime, end to end.

Camoufox 146.0.1
Product
RuntimeAction lockOptimised DOMMCP toolbeltVulpine-Box
Ecosystem
Foxbridgevulpine-markmobilebridgeOpenClawCamoufox
Resources
Resources hubRoadmapChangelogBlogDocs
Community
SupportDiscussionsContributingIssuesSecurity
© 2026 VulpineOSBuilt on Camoufox · Firefox 146.0.1TermsPrivacyCookiesAcceptable useSecurity