Responsible disclosure

Email [email protected]. PGP encryption is available — request our public key at the same address. Our machine-readable contact is published at /.well-known/security.txt.

A useful report includes:

• The affected host, endpoint, or component, and the vulnerability class.
• Reproduction steps or a proof of concept a reviewer can follow.
• The impact you believe it has, and any prerequisites (authenticated? which role? which plan?).
• Timestamps and source IPs of your testing, so we can separate it from real attacks.
• Whether you intend to publish, and on what timeline.

Please do not include third-party data in your report. If you accessed data that is not yours, stop, tell us in the report, and delete your copy.

• The Sentrail web application and its authenticated surfaces.
• The Sentrail API, server functions, and the MCP server, including token handling.
• Authentication, session, and authorization flows, including tenant isolation and row-level security bypasses.
• The agent runtime and approval gates — in particular anything that would let an action reach a production system without approval.
• Provider integrations (GitHub, Supabase, Vercel) and the storage and handling of their tokens.
• The static-analysis scanner and pentest runner services, including container isolation.
• Webhook endpoints and their signature verification.
• Billing flows where a flaw grants entitlements without payment.

• Automated scanner output without a demonstrated, working proof of concept.
• Denial of service through volumetric traffic, load testing, or resource exhaustion.
• Social engineering, phishing, or physical attacks against ZerwOne personnel or facilities.
• Missing hardening headers or best-practice recommendations with no demonstrated impact.
• Vulnerabilities requiring a rooted or malware-compromised device, or a heavily outdated browser.
• Issues in third-party services we do not operate — report those to the vendor.
• Email configuration findings (SPF, DKIM, DMARC) with no demonstrated spoofing impact.
• Self-XSS, clickjacking on pages with no sensitive action, or content spoofing without a security consequence.

• Test only against accounts and data you own. Create your own workspace rather than touching another tenant.
• Stop as soon as you confirm a vulnerability. Do not pivot, escalate persistence, or enumerate further than needed to prove impact.
• Do not access, modify, exfiltrate, or destroy data belonging to anyone else, and do not degrade the Service for other users.
• Do not use the scanner or pentest runners as a springboard against third-party systems.
• Give us a reasonable opportunity to fix the issue before disclosing it publicly.
• Comply with applicable law throughout.

If you make a good-faith effort to comply with this policy during your research, we will consider your activity authorized. ZerwOne Network LLC will not initiate or support legal action against you under computer-misuse law, breach of contract, or the anti-circumvention provisions of the DMCA in connection with that research, and we will make it known that your actions were authorized if a third party brings a claim.

This safe harbor covers only claims that ZerwOne Network LLC controls. It cannot bind our infrastructure providers or other third parties. If you are unsure whether a specific test is in scope, ask us first at [email protected] — we would much rather answer a question than argue about it afterwards.

We do not currently run a paid bug bounty programme, and we will not imply one exists. With your permission we credit reporters in our release notes and maintain a researchers page. If we launch a bounty, we will announce it here first.

This policy covers Sentrail itself. If you found a vulnerability in an application that happens to use Sentrail, report it to that application's owner, not to us. If you believe Sentrail is being used to scan a system without authorization, email [email protected] — our Acceptable Use Policy explains how we handle that.