Ward³
Back to overview

Security

Last updated: 2026-05-22

Ward³ builds adversarial-grade detection software. We hold ourselves to the same standards we ask of our customers. This page covers responsible disclosure for the Ward³ platform and this website, plus a summary of the security posture of the platform.

Reporting a vulnerability

If you believe you have found a security issue in the Ward³ platform, this website, or any related artifact, please email security@ward3.ai. Use PGP if you can — our key fingerprint is published at /security/pgp.txt. We aim to acknowledge reports within 2 business days and to triage them within 5. We do not pursue legal action against good-faith researchers who follow this process and do not exfiltrate or alter data beyond what is necessary to demonstrate the issue.

Scope

In scope: ward3.ai, the Ward³ open reference implementation, signed release artifacts, and any service explicitly operated by Ward³. Out of scope: third-party services we link to, customer-operated deployments of the Ward³ platform (please report through the customer's security team), and findings that require physical access to production hardware.

Platform security posture

The Ward³ platform ships with built-in integrity, governance and audit guarantees: cryptographically signed inference artifacts verified at load, append-only model registry with full provenance, hash-chained audit ledger signed with NIST-aligned post-quantum primitives, multi-party quorum (Shamir secret sharing) for high-impact actions, and a 4-eyes War Mode ceremony for line-rate enforcement actions.

Supply chain

Release artifacts are produced through reproducible builds and attested with Sigstore. The model registry tracks training dataset hash, training run identifier, and code commit for every deployed inference artifact. Software bill of materials (SBOM) is published with each release.

Cryptographic primitives

Post-quantum signatures (ML-DSA, NIST-aligned) are used end-to-end on the audit ledger and quorum operations. Post-quantum key encapsulation (ML-KEM) is used for confidential channels. Classical primitives remain in place where mandated by ecosystem compatibility (TLS, JWT).

Contact

Security: security@ward3.ai. PGP key: /security/pgp.txt. For non-security inquiries please use the Request access form.