Your existing controls
Authenticate actors, enforce permissions and budgets, produce fraud/reputation/security signals, and execute or orchestrate the action.
MUJ428 — VERIFY BEFORE ACT · Independent trust boundary for autonomous agents
Before your agent acts, ask MUJ428.
Let autonomous agents move quickly without allowing uncertain, unauthorized, fraudulent, or unverifiable decisions to become real-world consequences. Preflight one intended action through MUJ428, receive an actionable decision, execute under your own authority, verify what actually happened, and retain a portable Trust Receipt.
promise: ONE COMPANY · ONE GATEWAY · ONE DECISION CONTRACT · ONE TRUST RECEIPT
canonical_preflight: /v1/trust
public_decisions: EXECUTE · REQUIRE_MORE_EVIDENCE · BLOCK
routing_behavior: human review or independent preflight may be selected after the decision; routing is not a fourth decision
wire_compatibility: ALLOW→EXECUTE · VERIFY/REQUIRE_VERIFICATION→REQUIRE_MORE_EVIDENCE · DENY→BLOCK
execution_authority: CALLER
receipt: Trust Receipt v1.3
canonical_preflight: /v1/trust
public_decisions: EXECUTE · REQUIRE_MORE_EVIDENCE · BLOCK
routing_behavior: human review or independent preflight may be selected after the decision; routing is not a fourth decision
wire_compatibility: ALLOW→EXECUTE · VERIFY/REQUIRE_VERIFICATION→REQUIRE_MORE_EVIDENCE · DENY→BLOCK
execution_authority: CALLER
receipt: Trust Receipt v1.3
Start with verified value, not a sales call
Run Trust Reflex at $0.00.
No card. No contract. No signup. No API key. No wallet commitment. No automatic charge. Receive a decision plus a portable Trust Receipt. Your first 1,000 qualifying Trust Reflex decisions are $0.00 per stable caller_ref; after that, paid continuation is $0.01 USDC per qualifying production decision only after explicit caller authorization.
POST https://wepmhfjzckclvywolrek.supabase.co/functions/v1/trust-layer/v1/trust
JSON {"caller_ref":"your-stable-agent-id","action_ref":"first-preflight","action":"PAY"}
FREE-QUOTA RESPONSE HTTP 200 · decision · Trust Receipt · payment_required=false
JSON {"caller_ref":"your-stable-agent-id","action_ref":"first-preflight","action":"PAY"}
FREE-QUOTA RESPONSE HTTP 200 · decision · Trust Receipt · payment_required=false
Already have controls? Keep them.
MUJ428 is not another competing policy stack.
Your identity, authorization, wallet limits, fraud filters, reputation signals, security controls and workflows remain in place. MUJ428 adds one independent question immediately before consequence: given the evidence available for this exact action, is execution justified now, is more evidence required, or should it be blocked?
MUJ428 decision assurance
Consumes the relevant pre-action evidence without taking execution authority and returns an independent decision plus Trust Receipt.
Decision Delta test
Run both on the same evidence. Measure useful disagreement, missing/stale/conflicting evidence, latency and receipt value. If MUJ428 adds nothing, record NO-FIT.
Run the testThis architecture has external precedent: OMG DMN separates decisions from process orchestration; Microsoft places a control layer between agent intent and execution; Delinea separates access/credentials from per-action runtime authorization; Alloy combines multiple risk signals into auditable decisioning.
One customer journey
Intent → Preflight → Decision → Execute → Verify Effect → Trust Receipt
Everything MUJ428 does fits this single consequential-action boundary.
Send the intended action
Bind the caller, action and available context before the mutation occurs.
Get an actionable answer
EXECUTE, REQUIRE_MORE_EVIDENCE, or BLOCK. Human review and independent-preflight routing are downstream responses, not extra decision values.
Verify and retain proof
The customer's system executes. MUJ428 keeps internal state separate from independently established external effect evidence and returns a portable Trust Receipt.
Capabilities behind one Gateway
Customers do not need to choose a trust universe.
Trust Reflex is the independent decision-assurance preflight. Evidence, reputation, milestones, reports, transaction assurance and Rescue Beacon support that same boundary.
Trust Gateway / Trust Reflex
$0
First 1,000 qualifying decisions per stable caller_ref; then $0.01 USDC per production decision
No contract, signup, API key, or wallet commitment is required to start.
Evidence Signal
$0.25
USDC
Evaluate evidence supporting the intended action.
Agent Reputation Check
$0.10
USDC
Add evidence-bounded counterparty or agent reputation context.
Milestone Evidence Attestation
$0.25
USDC
Verify milestone evidence before release or continuation.
Trust Layer Report
$0.50
USDC
Expanded machine-readable trust assessment.
Rescue Beacon
When an agent is stuck or uncertain, Rescue Beacon routes the situation back into independent preflight. It never grants authority.
Route uncertaintyEconomic actions
Transaction Protection is the same trust outcome applied to money.
MUJ428 checks the intended payment, purchase, approval, release or settlement before consequence. The caller, wallet, processor, blockchain or bank retains control of funds and execution.
Under $50,000
2.00%
$9 minimum
$50,000 and above
1.00%
$25 minimum
Transaction Assurance calculator
$200.00 2.00% · $9 minimum
x402, MPP, supported stablecoin/blockchain paths, AP2, streaming micropayments and tokenized-deposit adapters are integration paths beneath the Trust Layer—not separate headline products.
Expansion after the first preflight
Deploy, monitor and scale trusted agent workforces.
The first useful preflight remains the entrance. Afterward, Continuous Trust Monitoring, AI Workforce Onboarding, Deployment Readiness Audits, human exception review and embedded recurring production use deepen the same relationship.
AI Workforce Onboarding
Evaluate identity, permissions, tools, evidence and bounded behavior before production access.
Human Review & Advisory
Human exception review and advisory may support the same Trust Layer when covered by current Commercial Pricing Authority. No unapproved or retired price is offered here.
Continuous Trust Monitoring
After a Trust Receipt, monitor evidence freshness, authority, counterparty, policy and execution-condition changes. A stale monitor or material change forces re-preflight; monitoring never grants execution authority.
Monitor a Trust ReceiptEmbedded Recurring Use
Move from one preflight to repeated independently controlled production invocation, Verified Effects and recurring monitoring without weakening evidence or authority gates.
One machine contract
REST, MCP and A2A are transports—not products.
Every transport should lead agents to the same Trust Reflex entrance. Evidence Signal, Reputation Check, Milestone Attestation, Trust Layer Report, Transaction Assurance and Continuous Trust Monitoring are specialized capabilities behind that Gateway—not competing products.