Skip to main content
Proof of Human Intent

PoHI for Autonomous Finance

When agents can move money, authentication is not enough. PoHI makes human authority a cryptographic primitive: atomic, auditable, enforceable.

Atomic authorizationModel-agnosticWorks with agent kits and smart accounts
Designed for irreversible boundaries, high-frequency execution, and verifiable accountability.
The missing layer
Between agent execution and financial consequence, enforceable human authority is not optional.
Agent
Agent Kit
PoHI Control Plane
Smart Account
Execution
Guarantee
No proof, no execution
Outcome
Audit-grade authority
Benefit
Lower attack surface
Point of no return
In irreversible systems, intent must be proven before execution, not investigated after.

Autonomy meets irreversibility

Crypto actions are global and final. Agents multiply execution speed and frequency. The result is a structural gap between identity and authority.

Agents outpace supervision

Delegated agents execute continuously. Humans cannot review every action, especially under market pressure and high-throughput conditions.

MFA proves identity, not intent

Today’s flows confirm a session or a device, but rarely produce a cryptographic proof that a specific human intended a specific action at the final boundary.

Bundled stacks create seams

Multiple systems and time steps introduce UI deception, session abuse, policy drift, and ambiguous audit trails across vendors and components.

Execution is scaling. Authority infrastructure has not caught up.
PoHI is designed for the last irreversible boundary.

Bundle vs primitive

Modern systems approximate intent by stacking signals. PoHI collapses authentication, confirmation, and authorization into a single atomic cryptographic act.

Today: intent as a bundle
  • MFA / 2FA + confirmation screens + policy engines
  • Multiple trust boundaries and timestamps
  • Non-atomic, integration seams, replay and race exposure
  • Audit trails are fragmented across systems
Intent is often inferred, not proven.
PoHI: intent as a primitive
  • One act → one proof → one enforcement point
  • Cryptographically binds human + action + constraints + time
  • Atomic: either proof exists or execution is blocked
  • Simplifies backend heterogeneity and reduces attack surface
Intent becomes verifiable and enforceable.
Definition
PoHI is a cryptographic proof that a specific human intentionally authorized a specific action, under explicit constraints, at a specific moment.
Atomicity
No proof, no execution.
Non-repudiation
Audit-grade evidence of authority.
Delegation
Bounded autonomy, revocable by design.

How it works

A simple pre-execution gate for high-risk actions, compatible with agent kits, smart accounts, and programmable money.

01
Agent proposes an action
Trade, transfer, policy change, withdrawal, freeze or override.
02
Constraints are computed
Scope, limits, time windows, tool permissions, destinations, thresholds.
03
User performs PoHI
One atomic interaction on device at the final decision boundary.
04
PoHI proof is produced
A verifiable artifact binding human + action + constraints + time.
05
Execution is enforced
Smart account / backend verifier checks proof before the irreversible step.
Key property
PoHI is a control plane, not an agent framework. It is model-agnostic and designed to be the final authority gate before irreversible execution.

Integrations

Fit existing stacks with minimal disruption: add a pre-execution PoHI gate to the actions that matter.

Agent frameworks

PoHI becomes the approval gate for high-risk tools. Agents can operate autonomously within pre-approved constraints, and stop automatically outside the boundary.

Tool routerPolicy boundAudit trail
Smart accounts and wallets

Attach PoHI proofs to critical operations: transfers, approvals, policy changes, withdrawals, and session delegation. Verification happens at the final execution point.

Account abstractionSession controlNon-repudiation
Stablecoins and compliance hooks

Bind human authority to freeze/unfreeze, overrides, recalls, and emergency actions with verifiable mandates and auditable delegation constraints.

MandatesOverridesEmergency control
Integration surface
Client
PoHI SDK for atomic user interaction
Verifier
Backend / smart account verification before execution
Policy
Optional constraints: scope, limits, time windows, thresholds
Audit
Single proof format for reporting and internal review

Use cases

PoHI is designed for high-impact actions: when a mistake, compromise, or agent overreach becomes immediate loss or liability.

For exchanges and platforms
  • Agent trading with bounded authority and hard stops
  • Human-approved withdrawals and whitelist changes
  • Session delegation with explicit limits and expiry
For stablecoin issuers
  • Authorized freeze/unfreeze and emergency overrides
  • Mandates and delegation for compliance operators
  • Audit-grade proof of legitimacy for sensitive actions
For chains and ecosystems
  • Standard authority primitive for agentic applications
  • Governance legitimacy for emergency controls
  • Composable “intent receipt” for critical state changes
Consistent principle
Whether transactions are irreversible or partially reversible, authority must be provable. PoHI binds execution and overrides to explicit human mandates.

Security posture

Atomicity simplifies the system. Simplification reduces failure modes, integration seams, and attack surface.

Why atomic reduces attack surface
  • Single enforcement point instead of multi-step flows
  • Less state to exploit: fewer sessions, fewer transitions
  • Less vendor and component heterogeneity in the approval path
  • Clear, verifiable artifact for audits and incident response
Threat classes PoHI targets
Session hijackUI deceptionAgent overreachPolicy bypassReplay ambiguityOverride legitimacy

PoHI is designed to be checked at the final boundary before execution, where authority must be strongest and most explicit.

Practical outcome
PoHI improves user experience and security simultaneously by collapsing multiple approval steps into one atomic interaction and a single verification path.

Roadmap for a pilot

A lightweight, practical path: prove value on 1–2 high-risk actions, then expand to delegation and policy templates.

Step 1

Add a PoHI gate to one irreversible action (withdrawal, transfer, trade execution, freeze/override).

Fast validation, immediate security impact.
Step 2

Introduce delegation constraints: scope, limits, time windows, tool permissions, thresholds.

Bounded autonomy for agents and operators.
Step 3

Standardize an audit-ready “intent receipt” format for internal review and reporting workflows.

Operational clarity and compliance alignment.

Start with a focused integration

We can scope a pilot around a small number of high-risk actions and ship an end-to-end PoHI proof and verifier path with minimal disruption to your stack.

contact@interstellar.network -Calendly link coming soon-
What you get in a pilot
  • PoHI gate for one high-risk action
  • Verifier integration at the final boundary
  • Initial policy template (constraints, limits, time)
  • Audit-grade proof artifact
Design principle
Atomic authority
Integration posture
Minimal disruption
Target outcome
Provable legitimacy