Report a vulnerability

Found a security vulnerability in Corgent? Thank you. Here is how to report it and what you can expect from us.

As of 13 October 2026

How to report

Email security@corgent.ai in German or English. Tell us what is affected (URL, app version, endpoint), how to reproduce it, the impact you see and whether you think it is already being exploited. We do not publish a PGP key yet; if you need encryption, write first without details and we will agree on a channel.

security@corgent.ai

Scope

In scope: corgent.ai, app.corgent.ai, api.corgent.ai, mcp.corgent.ai, the Corgent app for Android and iOS, the embeddable chat widget, and the MCP servers and SDKs we publish.

Out of scope: denial-of-service and load tests, social engineering and phishing, physical attacks, third-party services (please report to them directly), findings that need a rooted device or a man-in-the-middle without a further flaw, missing headers without a demonstrated impact, and prompt injection that only changes what the model tells the person who injected it. Prompt injection that crosses a tenant, approval or data boundary is explicitly in scope.

What you can expect

  • Acknowledgement within 2 working days.
  • First assessment (valid, severity) within 5 working days.
  • Fix: critical within 7 days, high within 30 days, medium and low within 90 days.
  • Coordinated disclosure after the fix, at the latest 90 days after the report, agreed with you.
  • We keep you informed and, if you like, name you in the acknowledgements. We do not pay bounties at the moment.

If a vulnerability is actively exploited, we report it under Art. 14 of the Cyber Resilience Act to the competent CSIRT through ENISA's reporting platform and inform affected users.

What we expect from you

  • Act in good faith and only as far as needed to show the vulnerability.
  • Use your own test accounts. Do not access, change or delete data of other customers. If you reach personal data by accident, stop, keep no copy and tell us.
  • No denial of service, no spam, no social engineering.
  • Give us reasonable time to fix before you publish.

Safe harbour

If you follow this policy, we consider your research authorised, will not file a criminal complaint or take civil action against you for it, and will say so to third parties. We cannot speak for third parties or public prosecutors.

Acknowledgements

No reports with consent to be named have been received so far.

security.txtTrust Center