How FLINT Speaks MITRE F3

FLINT records are built for a practical handoff. A merchant, risk team, fraud analyst, compliance officer, or partner system should be able to inspect a signed verification record and understand what FLINT observed before an agent-driven transaction moved.

MITRE F3 gives that handoff a common vocabulary. FLINT does not claim MITRE endorsement or certification. The correct framing is narrower and more useful: FLINT verification records are interpretable as MITRE F3 sightings.

What F3 Adds

F3 is a public fraud framework from MITRE's Center for Threat-Informed Defense. It gives fraud behavior a structured set of tactics and techniques. That structure is useful because agent-commerce events will cross merchants, rails, wallets, runtime environments, and investigation teams.

FLINT does not replace its native KYA layers with F3. The six FLINT layers remain Principal Identity, Agent Identity, Wallet Provenance, Authorization Scope, Environment Identity, and Cross-Merchant Reputation. F3 sits beside those layers as a label set that helps downstream teams interpret what a record observed.

That makes F3 additive. A FLINT record still answers whether the agent was known, authorized, in scope, running in a trustworthy environment, and supported by reputation. F3 helps label the fraud technique when one of those signals maps cleanly to the public framework.

The Honest Labeling Rule

Phase 0 has one rule: label only what FLINT genuinely detects. A clean transaction carries no F3 indicators. A suspicious transaction carries only the technique identifiers tied to risk rules that actually fired.

Today, the public live mapping is intentionally small. When FLINT's Environment Identity layer sees suspicious runtime evidence consistent with device tampering or bot signals, the signed record can label F1023 Device Fingerprint Spoofing.

No other technique is emitted today. That restraint is the point. A trust company should be more embarrassed by over-claiming than by publishing its blind spots.

The Current Crosswalk

Principal Identity currently has no emitted F3 technique. Blind spots include F1031 Impersonate Account Holder and F1032 Impersonate Official.

Agent Identity currently has no emitted F3 technique. Blind spots include F1004 Access with Stolen Session Cookie and F1006.001 Account Takeover: Exposed API Key.

Wallet Provenance currently has no emitted F3 technique. Blind spots include F1006 Account Takeover, F1018 Convert to Cryptocurrency, F1045 Structuring, and F1025 Electronic Funds Transfer.

Authorization Scope currently has no emitted F3 technique. F1046 Test Payment Thresholds is a blind spot until the signal is specific enough to avoid over-labeling ordinary scope failures.

Environment Identity detects F1023 Device Fingerprint Spoofing when the runtime signal supports it. F1030 Geolocation Spoofing and F1007 Adversary-in-the-Browser remain blind spots.

Cross-Merchant Reputation is reserved for later labels once partner outcome data provides enough evidence to support them.

How Labels Enter the Record

F3 indicators live in the signed verification record under signals.mitre_f3.indicators. The record also carries a framework reference in envelope.framework_references, so downstream consumers can tell which public vocabulary is being used.

Every technique ID is checked against FLINT's bundled F3 data before it is surfaced. The public product does not render arbitrary technique IDs from an untrusted pasted record.

Outcome feedback adds the second half of the flywheel. A merchant can submit confirmed or rejected F3 ground truth after the transaction outcome is known. The Trust Graph can then connect the original signal, the technique label, and the confirmed outcome over time.

Why Blind Spots Are Visible

Public blind spots make the product more credible, not less. They tell buyers where FLINT has evidence today and where it does not.

They also create a clean roadmap. If a future partner needs coverage for a technique FLINT does not yet detect, the work begins with the signal, the data source, and the validation rule. The label comes last.

That discipline keeps F3 useful. It prevents a public vocabulary from becoming marketing decoration.

What Is Out of Scope

This article does not describe FLINT mitigation methods. Phase 0 is techniques and crosswalk only.

It also does not create an enterprise matrix dashboard, a TAXII feed, an FDE toolkit, or F3-aware MCP tools. Those belong to later phases and require different gates.

The durable move in Phase 0 is simpler: make every signed record and every outcome label capable of feeding a public, validated fraud vocabulary without pretending FLINT sees more than it actually sees.