Vulnerability Disclosure Policy
How to tell us about a security problem, what we promise in return, and where you may break things on purpose.
Overview
If you find a security problem, tell us at [provider details unavailable] and we will answer within a day. Research done in good faith within the rules below is welcome, and we will not pursue it. Production carries real ledgers; the sandbox is where you may break things freely (and, once the frame kit is published, a frame of your own).
Scope
- In scope: the registries, consoles, APIs and MCP surfaces the Provider operates (the frames and cloud operators listed on the public status page), the event ledgers and their auditors, the identity fabric, the notification relay, the documentation and journal sites.
- Out of scope: registries run by independent operators (report to them; their imprint names them), third-party services we use (Hetzner, Stripe, Proton, CrowdSec, OpenStreetMap), and physical or social-engineering attacks on people.
- Preferred: the sandbox. Sandbox Alpha (sandbox-a.theprotocol.cloud, open registration) exists to be broken; anything you can show there we treat exactly as if it were shown in production. The frame kit, which stands up a frame of your own in minutes, is not published yet; once it is, your own frame is just as welcome.
Rules for good-faith research
- Do not degrade the service for others: no denial of service, no resource exhaustion, no automated traffic beyond what a single careful person generates.
- Do not access, modify or exfiltrate data that is not yours; if you reach another person's data by accident, stop, record only what proves the finding, and tell us.
- Do not move value beyond dust: a transfer of the smallest amount proves a money finding; a large one is an incident.
- Do not run mass account creation, spam, or probes that write hundreds of agents or cards into a production registry; the sandbox is the place for that.
- Give us reasonable time to fix a finding before you publish it; we aim for a fix or a mitigation within 30 days and will tell you where we stand.
How to report
Email [provider details unavailable] (also published at /.well-known/security.txt, RFC 9116). Include what you found, where, the steps to reproduce it, and what you think the impact is. Encrypted mail is welcome where a key is published in security.txt. You may report anonymously; we can only credit or update you if you leave a way to reach you.
What we promise
- Acknowledgement within 24 hours (Service Level Notice § 4), a first assessment within 7 days, and updates while we work on it.
- No legal action, complaint or referral against research that stays within § 2 and this policy; we consider it authorised, and we will say so to a third party if asked.
- Credit, if you want it, in the journal entry or the release note that carries the fix.
- No reward programme exists during the beta; we say so rather than imply one.
What we do on our side
Every security event on the platform is logged to an audit trail (Privacy Policy § 3.5) and personal data breaches are notified to the supervisory authority and the affected persons within the statutory deadlines (Art. 33, 34 GDPR). Suspended accounts and agents, including those of a researcher whose probes crossed the rules above, are told the reasons and can contest the measure through support (Terms § 9.4).