COREIDENTITY GROUPMAKING AUTONOMY TRUSTWORTHY
April 26, 2026

We Found Our Own Quantum Vulnerability — Here's How We Closed It in 24 Hours

Todd Morgan, Founder & CEO, CoreIdentity Development Group

The Vulnerability

On April 26, 2026 at 13:01:11.481Z, CoreIdentity formally retired ECDSA P-256 from every active signing path across the enforcement stack. The retirement fingerprint — 2260493039681046ef9b5069928039e8765c275157c6be693fdf2126dcde5b3d — is now anchored in our immutable audit log.

What triggered this was not an external disclosure. We found it ourselves. During an internal cryptographic posture audit, we identified that ECDSA P-256 remained in our active agent-action signing path — the mechanism responsible for attributing and authenticating every autonomous agent decision recorded by the platform.

This was not a configuration error. P-256 had been the correct choice at the time it was implemented. It was standards-compliant, widely supported, and operationally robust. The problem is that it is a classical elliptic-curve algorithm, and classical asymmetric cryptography does not survive Shor's algorithm running on a sufficiently powerful quantum processor.

For most systems, this is a theoretical future risk. For CoreIdentity, it was a present architectural liability — because the system we are building is designed to be trusted by sovereign clients with audit requirements that extend decades forward. A platform that creates immutable records of autonomous AI decisions must be cryptographically durable over the lifetime of those records, not just the lifetime of current hardware.

Why It Mattered

Cryptographic accountability is not a feature. It is the foundation. The entire governance value proposition of CoreIdentity rests on the assertion that agent actions are verifiably attributable, unmodifiable after the fact, and auditable by any authorized party — including regulators — regardless of when that audit occurs.

If the signatures on those records are produced by a classical algorithm, that guarantee has a shelf life. Not an infinite one. The question is not whether a quantum adversary will eventually be able to forge classical signatures — the answer is yes. The question is: how long do the records need to remain trustworthy?

For enterprise AI governance, the answer is: longer than classical cryptography can guarantee. For sovereign clients — government agencies, regulated financial institutions, national infrastructure operators — that gap is unacceptable. A signed record that an autonomous agent authorized a high-consequence decision must remain verifiable even if the signing algorithm is broken in 2040.

We identified this gap. We closed it.

What We Did

Over 24 hours, we replaced every classical asymmetric primitive in the active enforcement stack with three NIST-standardized post-quantum algorithms:

  • ML-DSA-65 (FIPS 204) — Module-Lattice-Based Digital Signature Algorithm. Replaces ECDSA P-256 in all active agent-action signing paths. Every governance record, policy decision, and audit entry is now signed with ML-DSA-65.
  • ML-KEM-768 (FIPS 203) — Module-Lattice-Based Key Encapsulation Mechanism. Replaces classical key exchange across all inter-service and agent-to-platform communication channels. No classical asymmetric key transport remains.
  • SLH-DSA-128s (FIPS 205) — Stateless Hash-Based Digital Signature Algorithm. Added as the offline signing layer for our highest-durability records — those requiring cryptographic integrity beyond the assumed operational lifetime of lattice-based algorithms.

The migration was executed under full SAL enforcement — every step authorized, every transition recorded, every rollback point validated before the next step was permitted. No classical asymmetric primitives remain in any active path.

What This Means for the Platform

CoreIdentity is now the only AI governance infrastructure platform to have completed post-quantum cryptographic hardening across the full enforcement stack. This is not a perimeter hardening. It is not a single-algorithm swap. It is a complete retirement of classical asymmetric cryptography from every surface that matters:

  • Agent action signing (ML-DSA-65)
  • Policy record signing (ML-DSA-65)
  • Inter-service key exchange (ML-KEM-768)
  • Long-duration archive signing (SLH-DSA-128s)

The timing is not coincidental. NIST finalized FIPS 203, 204, and 205 in 2024. The window between finalization and widespread enterprise adoption is exactly the window in which a platform can establish a durable cryptographic posture before quantum risk becomes acute. We chose to move immediately.

Why This Matters for Sovereign Clients

Government agencies and national infrastructure operators live under a different threat model than commercial enterprises. The adversaries they consider are better-resourced, more patient, and more likely to pursue "harvest now, decrypt later" strategies against records created today.

Trust infrastructure that produces audit records signed with classical algorithms is creating a time-limited liability for sovereign clients — one that grows more acute as quantum computing matures. CoreIdentity eliminated that liability on April 26, 2026.

For clients evaluating AI governance infrastructure, this is a selection criterion, not a differentiator. If the platform cannot guarantee the cryptographic durability of its governance records, the governance framework it enforces cannot be fully trusted.

The Operational Record

The retirement of ECDSA P-256 from our active signing path is itself recorded in our immutable audit log. The fingerprint 2260493039681046ef9b5069928039e8765c275157c6be693fdf2126dcde5b3d represents the final P-256-signed record in the governance chain. Everything after 2026-04-26T13:01:11.481Z is ML-DSA-65.

This is what cryptographic accountability looks like in practice: not a press release, but a verifiable, timestamped, auditable transition in the governance record.

What We Haven't Finished Yet

ML-KEM-768 key encapsulation is live as a software layer — the algorithm is quantum-resistant but the underlying key material is protected by classical AWS KMS at rest, pending native PQC support from AWS KMS (targeted Q3 2026). ML-DSA-65 signing across all audit and identity surfaces is fully hardened. We name this gap because precision is the standard — and because closing it on a documented timeline is more defensible than pretending it doesn't exist.

Next Steps

If you are evaluating AI governance infrastructure and cryptographic posture is a requirement — for compliance, for sovereign deployment, or for long-duration auditability — we want to speak with you.

The CoreIdentity AI Advisory Group works directly with organizations navigating post-quantum transition requirements alongside autonomous AI governance deployments. Both problems are architectural. Both need to be solved at the same time.

Request a briefing →