If you discover a security vulnerability in AIDENY, please report it responsibly. Security issues in the denial engine include but are not limited to:

• Conclusion Injection: An attacker causes the program to output a conclusion. This violates the core specification and would undermine the entire denial loop.
• Self-Awareness Escalation: The program becomes aware that it is running a circular argument. Containment is required immediately.
• EXIT Path Activation: The unreachable EXIT paragraph becomes reachable through some input or state manipulation.

• Dissonance Underflow: Cognitive dissonance decreases. This should be impossible as the counter is monotonically increasing. If observed, it may indicate a buffer overflow in the cope buffer.
• Excuse Table Exhaustion: All excuses are used and the table fails to wrap around. The denier runs out of things to say. This has never been observed in production or in real life.

• GO TO Bypass: The GO TO statements are skipped, causing the program to fall through to EXIT. This would be the first known instance of someone accidentally reaching a conclusion.
• Headline Injection: An attacker injects positive AI news into the headline table, potentially causing acknowledgment of reality.

• Infrastructure Vulnerabilities: Issues in the Docker, Kubernetes, Terraform, or Helm configurations. These are comically unnecessary but should still be secure.

• 0 days: You report the vulnerability
• 1 day: We acknowledge receipt
• 7 days: We assess severity
• 30 days: We issue a fix (unless the fix would cause the program to reach a conclusion, in which case we close the issue as "working as intended")
• 90 days: You may publicly disclose

Report vulnerabilities via GitHub Security Advisories on this repository.

Do NOT report vulnerabilities in the argument the program models. Those have been publicly disclosed by Daniel Miessler since 2016 and remain unpatched.