AgentProof
AI compliance needs evidence. AgentProof helps teams document what an AI agent does, what it can touch, what evidence exists, and what still needs review before your compliance review — a compliance-readiness record, not a certification, not an audit, and not legal advice. It tracks named, public sources and turns them into practical, source-bound guidance.
Microsoft, NIST, ISO, OWASP, the EU, Google, the Cloud Security Alliance, and MITRE are publishing new guidance every quarter. AgentProof tracks those named sources and turns them into a practical compliance-readiness model — what to document, not a verdict on the law.
AgentProof's Capability Zones (Knowledge, Task / Tool-Assisted, Autonomous Action) plus five cross-cutting modifiers tell you which controls matter and which evidence a stakeholder will expect before your compliance review.
20 documented good-design patterns covering least-privilege tool access, human approval, audit trails, fallback paths, sensitive-data handling, and continuous reassessment.
20 documented anti-patterns linked to the named risks they create — including excessive permissions, broad autonomy without approval, and prompt-injection exposure.
Each readiness control names what acceptable, strong, and weak evidence looks like — so the record captures documented proof, and a stakeholder review is grounded rather than improvised.
The compliance-readiness record brings the whole story of an agent into one place — readiness, evidence, gaps, controls, and review flags — version-stamped. When the library evolves, AgentProof surfaces a calm advisory to reassess; it never silently re-scores.
AgentProof tracks watch-status signals (autonomy boundaries, identity, oversight, regulation, injection testing) and flags when an agent's record may need reassessment when supporting sources publish guidance.
Start with the Compliance Readiness Pilot — a fixed €1,500 engagement covering up to 5 AI agents — and build a versioned compliance-readiness record for each agent in your own workspace.
AgentProof's source-tracking + change-detection + proposed-update engine surfaces source changes. Nothing reaches scoring, the compliance-readiness record, or buyer-facing guidance without a recorded reviewer decision.
AgentProof compliance-readiness guidance is derived from named sources. Not a certification, audit, or legal compliance statement.
From the AgentProof Library
The public preview is free to view. The full library is available inside your workspace.
Explore capability zones →
See which capability zone your agent falls into and which controls apply.
Read good agent design →
Ten principles that separate a credible, documented agent from one that only works in an illustrative setting.
Preview the workspace experience →
See a sample of the deeper experience before signing in.