Add AI attribution to README and SECURITY docs

Adds a 'Built with Claude AI' badge to the README header and an
'AI Assistance' section acknowledging that the project was designed
and built with Claude by Anthropic. Updates SECURITY.md with an
'AI-Assisted Security Infrastructure' section detailing the specific
security work that was AI-generated — linting rules, unit tests,
CI pipeline, and credential handling design.

SECURITY.md

@@ -43,3 +43,14 @@ MikroTik Manager is designed to be run on your local network. A few things to ke
 - HTTPS is enforced by default. The included nginx configuration redirects all HTTP traffic to HTTPS.
 - Default credentials should be changed immediately after first login.
 - The built-in role-based access control (Admin / Operator / Viewer) should be used to limit user privileges.
+
+## AI-Assisted Security Infrastructure
+
+The security measures in this project were designed and implemented with the assistance of [Claude](https://claude.ai) by Anthropic, including:
+
+- **ESLint security rules** — static analysis configured with `eslint-plugin-security` to catch common vulnerabilities (injection risks, unsafe regex, insecure randomness, etc.)
+- **Unit tests** — test coverage for security-critical code paths including AES-256-GCM encryption/decryption, JWT signing and verification, and role-based access control middleware
+- **CI pipeline** — automated linting, type-checking, dependency auditing (`npm audit`), and CodeQL analysis on every push and pull request
+- **Credential handling** — encryption-at-rest design for stored device credentials, JWT secret management, and `.env`-based secrets isolation
+
+AI-generated security infrastructure is reviewed and validated the same as any other code. If you discover a gap or vulnerability, please report it using the process above.