Trust & Compliance

Mastyf Trust Center

How Mastyf protects your data, isolates execution environments, and satisfies enterprise compliance standards.

Architecture

Zero-Trust Reference Monitor

The reference monitor operates as an independent out-of-band proxy. Compromise of the LLM or agent memory bus cannot bypass transport-level complete mediation.

Privacy

Zero Payload Retention

Mastyf Gateway evaluates payloads ephemerally in memory. Prompt text and tool arguments are never stored or transmitted to external servers without explicit policy configuration.

Cryptography

Ed25519 Execution Receipts

Every authorization decision generates a signed, non-repudiable cryptographic receipt, ensuring that all agent actions can be independently audited.

Compliance Framework Alignment

OWASP Top 10 for Agentic Applications (2026)

Complete

Directly maps to and neutralizes Goal Hijacking (ASI01), Tool Misuse (ASI02), Identity Abuse (ASI03), and Cascading Failures (ASI05) through execution-boundary reference monitors.

NIST Software Agent Security (2026)

Aligned

Addresses non-repudiation, granular delegation, and cryptographic auditability for autonomous software agents accessing tools and data stores.

SOC 2 & ISO 27001 Control Mapping

Audit Ready

Cryptographic Ed25519 execution receipts provide tamper-evident proof of authorization and execution certainty for third-party compliance auditors.

Rigorous Engineering Trust

What Mastyf Guarantees — and Does Not Guarantee

In an industry filled with vague claims of “100% unbreakable AI firewalls”, we publish our formal security invariants and our explicit scientific limitations side by side.

Architectural Guarantees

Under formal Complete Mediation Axioms (A1–A6) and reference monitor integrity:

Under complete mediation axioms (A1–A6), no tool call outside authorized capability scopes can be dispatched to backend infrastructure.
Any blocked or uncertain tool invocation produces strictly zero bytes transmitted across the tool transport wire.
Learned semantic models (Mastyf Guard) can revoke or escalate authority, but can never synthesize or grant permission (Afinal ⊆ Astruct).
Unconfirmed tool execution (ExecutionCertainty = UNKNOWN) cannot silently advance dependent workflow state (Theorem 3).

Scientific Boundaries & Scope Limits

Honest constraints documented in our published research:

Mastyf does not make an autoregressive Transformer intrinsically safe; it prevents compromised model outputs from executing privileged actions.
Semantic neural auditing is distribution-bounded (empirical FNR of 21.5% on in-scope manipulations prior to boundary sharpening) and subject to adversarial evasion without deterministic invariants.
Decentralized Information Flow Control (DIFC) tracks explicit data flow channels; it does not eliminate covert timing or semantic side-channels.
Protection guarantees apply exclusively to tool execution paths routed through the Mastyf reference monitor.