Quantum Shield Advisor

Inspiration

The progress toward cryptographically relevant quantum computers introduces a quiet but critical risk to today’s digital systems. Even before quantum machines are widely available, attackers can already perform harvest-now–decrypt-later (HNDL) attacks by collecting encrypted traffic and data today and decrypting it in the future. While standards bodies such as NIST and government mandates like CNSA 2.0 define post-quantum cryptography (PQC) algorithms and timelines, there is a major gap between these standards and their practical adoption in real systems. This gap inspired the creation of Quantum Shield Advisor.

What it does

Quantum Shield Advisor is an AI-powered advisor that analyzes real-world systems to identify cryptography vulnerable to quantum attacks and provides actionable, standards-aligned migration guidance. It ingests source code, configuration files, infrastructure artifacts, secrets, and documentation, correlates findings across these inputs, and detects usage of quantum-vulnerable algorithms such as RSA, ECDSA, and NIST P-256. The system then assesses quantum exposure, highlights harvest-now–decrypt-later risks, and recommends NIST-approved post-quantum alternatives such as ML-KEM, ML-DSA, and SLH-DSA, including hybrid transition strategies when appropriate.

How we built it

The project was built using Gemini 3 for its multimodal reasoning and long-context capabilities. Gemini analyzes multiple inputs simultaneously — including code, configurations, and documentation — to construct a unified cryptographic inventory of a system. A structured analysis pipeline identifies vulnerable primitives, evaluates quantum risk, and generates prioritized recommendations. The results are presented through an interactive dashboard with a clear quantum risk score, detailed findings, migration guidance, and a structured JSON report suitable for audits and automation.

Challenges we ran into

One of the main challenges was balancing accuracy with responsibility. Post-quantum cryptography is a high-stakes domain, so the system had to avoid false confidence while still providing decisive guidance. Another challenge was correlating cryptographic usage across different layers of a system and multiple file formats without producing noisy or misleading results. Designing outputs that are both technically precise and understandable to different audiences was also a key challenge.

Accomplishments that we're proud of

We are proud of building a tool that goes beyond simple scanning to deliver context-aware, system-level cryptographic analysis. Quantum Shield Advisor accurately detects quantum-vulnerable algorithms across diverse inputs, correlates them into a single risk model, and translates complex standards into clear, actionable migration steps. The ability to produce structured reports and code-level guidance demonstrates how AI can meaningfully support post-quantum readiness efforts.

What we learned

Through this project, we learned how deeply classical cryptography is embedded across modern systems and how difficult post-quantum migration becomes without holistic visibility. We also learned that AI is most effective in security engineering when it reasons across multiple modalities and layers, rather than treating files or components in isolation. Clear communication of risk is just as important as technical correctness.

What's next for Quantum Shield Advisor

Future work includes expanding support for additional architectures and deployment environments, deeper validation of hybrid cryptographic configurations, continuous audit capabilities to track readiness over time, and broader ecosystem integration with DevSecOps tooling. As post-quantum standards mature, Quantum Shield Advisor aims to remain a practical companion for teams navigating the transition to quantum-resistant systems.

Built With

• and-structured-json-output-react-(auto-generated-via-ai-studio-build-mode)-?-frontend-interface
• code-generation
• dark-mode-ui
• file-upload
• google-ai-studio-(gemini-3-family-?-gemini-3-pro-/-flash)-?-core-reasoning-engine
• images
• multimodal-file-analysis-(code
• pdfs
• responsive-dashboard
• voice)