chore: ai compliance

SECURITY.md

@@ -2,15 +2,15 @@
 
 ## Reporting Security Issues
 
-Thanks for helping make Donut Browser safe for everyone! ❤️
+Thanks for helping keep Donut Browser safe.
 
-I take the security of Donut Browser seriously. If you believe you have found a security vulnerability in Donut Browser, please report it to me through coordinated disclosure.
+I take the security of Donut Browser seriously. If you believe you have found a security vulnerability, report it to me through coordinated disclosure.
 
-**Please do not report security vulnerabilities through public GitHub issues, discussions, or pull requests.**
+Do not report security vulnerabilities through public GitHub issues, discussions, or pull requests.
 
-Instead, please send an email to **[contact@donutbrowser.com](mailto:contact@donutbrowser.com)** with the subject line "Security Vulnerability Report".
+Instead, send an email to [contact@donutbrowser.com](mailto:contact@donutbrowser.com) with the subject line "Security Vulnerability Report".
 
-Please include as much of the information listed below as you can to help me better understand and resolve the issue:
+Include as much of the following as you can:
 
 - The type of issue (e.g., buffer overflow, injection attack, privilege escalation, or cross-site scripting)
 - Full paths of source file(s) related to the manifestation of the issue
@@ -21,18 +21,26 @@ Please include as much of the information listed below as you can to help me bet
 - Impact of the issue, including how an attacker might exploit the issue
 - Your assessment of the severity level
 
-This information will help me triage your report more quickly.
+This helps me triage your report faster.
+
+## AI-Assisted Reports
+
+Use AI to find vulnerabilities. Fuzzing, static analysis, a model reading the code: all fine, and some of it works well.
+
+The report itself has to be written by a human, and verified by that human. Before sending, confirm the vulnerability exists in the current code, at the paths you cite, and that you can reproduce it. An unverified model-written report is not a security report; it will be closed without analysis.
+
+Say in your email whether AI was involved and what it did. That disclosure is never held against you. Omitting it is what ends the conversation.
 
 ## What to Expect
 
-- **Response Time**: I will acknowledge receipt of your vulnerability report within 72 hours.
-- **Investigation**: I will investigate the issue and provide you with updates on my progress.
-- **Resolution**: I aim to resolve critical security issues as fast as possible, but no longer than in 30 days after the initial report.
-- **Disclosure**: I will coordinate with you on the timing of any public disclosure.
+- Response Time: I will acknowledge receipt of your vulnerability report within 72 hours.
+- Investigation: I will investigate the issue and send you updates on my progress.
+- Resolution: I aim to resolve critical security issues as fast as possible, and no later than 30 days after the initial report.
+- Disclosure: I will coordinate with you on the timing of any public disclosure.
 
 ## Contact
 
-For urgent security matters, please contact me at **[contact@donutbrowser.com](mailto:contact@donutbrowser.com)**.
+For urgent security matters, contact me at [contact@donutbrowser.com](mailto:contact@donutbrowser.com).
 
 For general questions about this security policy, you can also reach out through: