SafeGate / Controlled Pilot Readiness

Historical Pilot Readiness Snapshot

Historical readiness snapshot: this page documents SafeGate before the completed real Base Mainnet merchant commerce proof. Current status is presented on the main Commerce Verification Hub. The current public position includes V9 payment-spine evidence, V9.1 safe negative backend validation, V13 controlled hardening scope, fee architecture, public validation, backend policy simulation, and a static agent-readable trust preview.

SafeGate verifies what happened after payment. SafeGate does not process payments, does not custody funds, and does not act as escrow.

Payment is the trigger. Trust is the product.

AI will automate payments. SafeGate will automate trust.

Payment Spine Passed Backend Validation Passed Public Surface Validation Passed Backend Policy Simulation Passed Agent-Readable Preview Open Pilot Planning Open V13 Hardening Open Fee Architecture Documented Real Backend Hardening Not Passed Yet Not Production Ready Not Formally Audited
Main Review Hub Pilot Review Index V13 Controlled Hardening Fee Architecture V13 Public Surface Validation V13 Backend Policy Simulation Agent-Readable Trust Preview Public Repository

Current Pilot Readiness Object

{ "project": "SafeGate", "pilotReadiness": "CONTROLLED_PILOT_PLANNING_OPEN", "coreLine": "SafeGate verifies what happened after payment.", "slogan": "Payment is the trigger. Trust is the product.", "agentVision": "AI will automate payments. SafeGate will automate trust.", "readyForReview": [ "V9_PAYMENT_SPINE_EVIDENCE", "V9_1_BACKEND_BEHAVIOR_VALIDATION", "V13_CONTROLLED_HARDENING_SCOPE", "FEE_ARCHITECTURE_DECISION", "V13_PUBLIC_SURFACE_VALIDATION_PASSED", "V13_BACKEND_POLICY_SIMULATION_PASSED", "AGENT_READABLE_TRUST_PREVIEW_OPEN" ], "notYetClaimed": [ "PRODUCTION_READY", "FORMAL_AUDIT", "OFFICIAL_PI_PARTNERSHIP", "PI_MAINNET_SETTLEMENT", "REAL_BACKEND_HARDENING_PASSED", "MCP_ENDPOINT_ENABLED", "TOOL_ENDPOINT_ENABLED", "AGENT_EXECUTION_ENABLED" ], "agentReadableTrustPreview": { "status": "STATIC_PREVIEW_ONLY", "mcpEndpointEnabled": false, "toolEndpointEnabled": false, "agentExecutionEnabled": false, "autonomousPaymentPermission": false, "autonomousAccessUnlockPermission": false } }

Current Pilot Readiness Position

Area Status
Controlled Pi Testnet Payment Spine PASSED
V9.1 Safe Negative Backend Validation PASSED
V10 Submission Ready READY FOR SERIOUS TECHNICAL REVIEW
V10.1 Review Feedback OPEN
Controlled Pilot Planning OPEN
V13 Controlled Hardening Scope OPEN
V13 Public Surface Validation PASSED 12/12
V13 Backend Policy Simulation PASSED 12/12
Agent-Readable Trust Preview OPEN / STATIC ONLY
Fee Architecture Decision DOCUMENTED
Real Backend Hardening Evidence NOT PASSED YET
MCP / Tool Endpoint NOT ENABLED
Agent Execution NOT ENABLED
Production Readiness NOT CLAIMED
Formal Third-Party Audit NOT CLAIMED
Official Pi Partnership NOT CLAIMED

What Is Ready

  • Public review hub is live.
  • Controlled Pi Testnet payment-spine evidence is available.
  • V9.1 safe negative backend validation evidence is available.
  • V10 submission-ready review position is documented.
  • V13 controlled hardening scope is open.
  • V13 Public Surface Validation passed 12/12.
  • V13 Backend Policy Simulation passed 12/12.
  • Fee Architecture Decision is documented.
  • Agent-Readable Trust Preview is open as a static public-safe preview.

What Is Not Yet Ready

  • SafeGate is not production-ready.
  • SafeGate is not formally audited.
  • V13 real backend hardening has not passed evidence validation yet.
  • Commercial pilot completion is not claimed.
  • Pi Mainnet settlement is not claimed.
  • Official Pi partnership is not claimed.
  • Native Pi split is not claimed if not supported.
  • Automatic fee deduction is not claimed where split does not exist.
  • MCP/tool endpoints are not enabled.
  • Agent execution is not enabled.

Controlled Pilot Entry Conditions

A controlled pilot should remain limited, public-safe, and evidence-driven.

Condition Expected State
Reviewer can understand what SafeGate does Clear positioning and public review hub available
Payment-state evidence exists Controlled Pi Testnet payment-spine evidence available
Negative backend checks exist V9.1 safe negative validation available
Hardening scope is clear V13 controlled hardening scope open
Public surface is consistent V13 Public Surface Validation passed 12/12
Policy behavior is simulated V13 Backend Policy Simulation passed 12/12 as client-side simulation
Fee model is not ambiguous 100 + 1 fee architecture decision documented
Agent-readable direction is bounded Static preview open; no MCP/tool endpoint and no agent execution
Boundaries are clear No custody, no escrow, no production-ready claim
Evidence is public-safe No passphrases, private keys, raw secrets, or sensitive data

V13 Must Improve Before Broader Pilot

  • Duplicate callback behavior
  • Idempotency
  • Replay resistance
  • Payment / invoice mismatch handling
  • Pi API timeout / ambiguous response behavior
  • Durable state failure behavior
  • Public verify unknown or mismatched pair behavior
  • Safe error handling
  • Access unlock regression
  • Fee settlement confirmation
  • Split / non-split payment handling

Agent-Readable Must Wait For Hardening

  • Agent-readable does not mean agent-executable.
  • No MCP endpoint before backend hardening.
  • No tool endpoint before backend hardening.
  • No autonomous payment permission.
  • No autonomous access unlock permission.
  • No receipt/evidence creation without verified final state.
  • No public verified outcome for unknown or mismatched pairs.
  • No private key or passphrase request.

Fee Readiness Rule

SafeGate may use a transparent 100 + 1 verification fee model.

  • Where native non-custodial split is supported, the fee can be routed automatically.
  • Where native split is not supported, SafeGate uses a separate verified fee payment.
  • SafeGate does not use postpaid merchant debt as the primary fee model.
  • SafeGate does not finalize verified receipt, evidence, access unlock, or public verify until required payment and fee states are confirmed.

Safe Agent Principles

  • Read before act.
  • Verify before trust.
  • Fail closed on ambiguity.
  • Never unlock before final state.
  • Never expose secrets.
  • Never treat simulation as real settlement.
  • Never expose MCP/tools before hardening.

Controlled Pilot Stop Conditions

A pilot should stop, pause, or remain limited if any critical trust boundary fails.

Stop Condition Reason
Access unlocks before backend-verified payment state Core trust rule broken
Receipt/evidence created before verified state False trust record risk
Unknown public verify pair returns active verified trust Public verification integrity risk
Replay or mismatch creates verified outcome Payment integrity risk
Timeout or ambiguous response is treated as verified Fail-secure boundary broken
Public output exposes secrets or sensitive data Security boundary broken
Fee-required flow finalizes without fee confirmation Fee settlement boundary broken
Agent can execute action before hardening Unsafe automation boundary broken

What Pilot Evidence May Include

  • Public-safe screenshots
  • Public-safe screen recordings
  • Test device/browser context
  • Expected vs observed behavior
  • Pass/fail count
  • Receipt/evidence/public verify status
  • Access locked/unlocked state
  • Fee settlement state where applicable
  • Agent-readable static preview status
  • Known limitations

What Pilot Evidence Must Not Include

  • Passphrases
  • Private keys
  • Wallet secrets
  • Service role keys
  • Raw payment secrets
  • Raw database errors
  • Raw Pi API responses
  • Sensitive customer data
  • Non-public personal information

Reviewer Path

Step Link Purpose
1 Main Review Hub Start with the current public position.
2 Pilot Review Index Follow the recommended review order.
3 V9 Payment Spine Evidence Inspect the core controlled payment-spine evidence.
4 V9.1 Backend Validation Inspect safe negative backend validation behavior.
5 V13 Controlled Hardening Review hardening scope and pass gate.
6 Fee Architecture Review fee model, split/non-split handling, and no-custody boundary.
7 V13 Public Surface Validation Review public page consistency and claim-boundary validation.
8 V13 Backend Policy Simulation Review client-side policy simulation and its limits.
9 Agent-Readable Trust Preview Review the static public-safe future direction for app and AI-agent-readable trust states.

Main Pilot Question

Can SafeGate maintain safe post-payment trust behavior when a real pilot introduces duplicate callbacks, replay attempts, mismatches, timeouts, durable state failures, fee-settlement requirements, public verification pressure, and future agent-readable trust-state consumers?

Pilot readiness depends less on happy-path success and more on safe behavior during failure, mismatch, replay, timeout, incomplete-state, fee-required, and automation-pressure conditions.

Current Safe Statement

SafeGate is ready for serious technical review and controlled pilot planning.

SafeGate is not production-ready, not formally audited, and not claiming official Pi partnership.

V13 Controlled Hardening Scope is open. Fee Architecture Decision is documented. Public surface validation and backend policy simulation have passed as public-safe checks. Agent-readable trust preview is open as a static preview only.