Built by security professionals,
for security professionals
IDProva was created because we saw a gap: AI agents are being deployed into critical infrastructure with no standardized way to verify their identity, scope their permissions, or audit their actions.
Tech Blaze Consulting
Canberra, ACT, Australia
Pratyush Sood
Founder & Principal Consultant
- ✓ ASD-endorsed IRAP Assessor
- ✓ Australian Government security cleared
- ✓ 20+ years in IT & cybersecurity
- ✓ Current ASD contract holder
Credentials
- ✓ NIST NCCoE — AI agent identity submission
- ✓ IRAP assessments for cloud providers
- ✓ Essential Eight maturity assessments
- ✓ DISP compliance advisory
NIST NCCoE Engagement
IDProva was submitted to the NIST National Cybersecurity Center of Excellence as part of their work on AI agent security. The protocol directly addresses the identity, delegation, and accountability gaps identified in the NIST AI Risk Management Framework.
Open Source Commitment
The IDProva protocol, core library, CLI, SDKs, and self-hosted registry will always be Apache 2.0 open source. We believe the security infrastructure for AI agents should be transparent, auditable, and free from vendor lock-in.
IDProva Cloud is a managed infrastructure layer on top of the same open-source code. If you stop paying, you keep running — just point at your own registry.
Security Posture
- ✓ 247 tests across 5 Rust crates — memory-safe by default
- ✓ Ed25519 (EdDSA) only — no algorithm negotiation vulnerabilities
- ✓ BLAKE3 hash chains — tamper-evident audit trails
- ✓ Parameterized SQL only — no injection by design
- ✓ Zeroize on drop for all key material
- ✓ STRIDE threat model documented
- ✓ SECURITY.md — responsible disclosure policy