Audits & Security

Operational security posture

We publish the ecosystem's security posture openly. Audits are continuous, execution is bounded, and the platform is non-custodial by design.

Six guarantees

Non-custodial architecture

Keys never leave the user's wallet. The platform cannot move user assets without an explicit in-wallet signature.

Bounded execution

Every AI agent operates within enforced execution bounds. Agents recommend; only signed user transactions execute on-chain.

Reversible where possible

All adaptive refinements are preview-before-apply with a Revert option. On-chain actions remain final by nature.

AI-assisted audit logic

Contracts entering the ecosystem are pre-screened by interpretable AI auditors that flag risk patterns for human review.

Contract validation

Pool contracts use validated, common patterns. Custom logic is held to higher review thresholds before ecosystem inclusion.

Closed alpha disclosure

We are in disciplined closed alpha. Surface area is intentionally bounded. Treat all participation as alpha-grade.

Active reviews

SurfaceStatusWindow
Core platform contractsin calibrationQ2 2026
DEX routing surfacein calibrationQ2 2026
Referral attribution layerdesign reviewQ3 2026
Bounded execution governanceinternal reviewQ2 2026

Detailed reports will be linked here as each review window completes.

Responsible disclosure

If you discover a vulnerability, please disclose it privately through the ecosystem channel before public posting. The closed-alpha posture allows us to address issues quickly without putting participants at risk.