Proof Verification
VORT generates cryptographic proofs for every intent execution, enabling trustless verification that agents behaved correctly. This guide explains how to understand, verify, and use these proofs.
The Proof Explorer
The Proof Explorer is VORT's interface for interacting with execution proofs. It provides tools for viewing, analyzing, and verifying proofs without requiring deep cryptographic expertise.
Accessing Proofs
Proofs can be accessed through multiple interfaces:
Dashboard Interface: The VORT Dashboard provides a visual Proof Explorer where users can view proofs for their intents:
Navigate to an intent in the dashboard
Click "View Proof" to open the Proof Explorer
The explorer displays proof structure, verification status, and detailed breakdown
SDK Interface: Programmatically access proofs through the SDK:
const receipt = await vort.getExecutionReceipt(intentId);
const proof = receipt.proof;API Interface: Direct API access for integration:
curl https://api.vort.protocol/v1/intents/{intentId}/proofBlockchain Interface: Proofs are also stored on-chain for permanent verification:
Proof hashes are stored in on-chain receipts
Full proofs can be retrieved from IPFS or other decentralized storage
On-chain storage ensures proofs are permanently available
Proof Structure
Each proof contains multiple components:
Proof Header: Metadata about the proof:
Proof type (ZK proof, Merkle proof, signature proof)
Generation timestamp
Intent identifier
Agent identifier
Verification key reference
Proof Data: The actual cryptographic proof data:
Encoded proof bytes
Public inputs
Verification parameters
Supporting evidence
Attestation Data: Attestations of specific execution aspects:
State transition attestations
Constraint satisfaction attestations
Outcome alignment attestations
Agent behavior attestations
On-Chain Evidence: References to on-chain data:
Transaction hashes
Block numbers
State root hashes
Event logs
Visual Proof Representation
The Proof Explorer provides visual representations of proofs:
Proof Tree: Hierarchical visualization of proof structure, showing how micro-proofs aggregate into intent-level proofs.
Execution Timeline: Timeline view showing when each proof component was generated during execution.
State Transitions: Visual representation of state changes, showing before/after states with proof attestations.
Constraint Verification: Visual indicators showing which constraints were verified and their verification status.
Agent Actions: Breakdown of agent actions with proof attestations for each action.
ZK Attestation Fields
VORT uses zero-knowledge proofs to enable verification without revealing sensitive information. Understanding the attestation fields is crucial for effective proof verification.
State Transition Attestations
State transition attestations prove that on-chain state changed correctly:
Before State Hash: Cryptographic hash of the state before execution, enabling verification that execution started from the expected state.
After State Hash: Cryptographic hash of the state after execution, enabling verification that execution ended in the expected state.
Transition Proof: ZK proof that the transition from before to after state was valid according to the intent requirements.
State Diff: Attestation of the specific state changes that occurred, proving that only intended changes happened.
Example:
Constraint Satisfaction Attestations
Constraint satisfaction attestations prove that user constraints were respected:
Slippage Attestation: Proof that execution slippage was within declared limits:
Declared max slippage: 0.5%
Actual slippage: 0.3%
Attestation: ✓ (within limits)
Cost Attestation: Proof that execution costs were within declared limits:
Declared max gas cost: 0.01 SOL
Actual gas cost: 0.008 SOL
Attestation: ✓ (within limits)
Protocol Attestation: Proof that execution used only allowed protocols:
Allowed protocols: ['Jupiter', 'Drift']
Protocols used: ['Jupiter', 'Drift']
Attestation: ✓ (only allowed protocols used)
Timing Attestation: Proof that execution completed within time constraints:
Deadline: 2024-01-01 12:00:00
Completion time: 2024-01-01 11:58:30
Attestation: ✓ (completed before deadline)
Outcome Alignment Attestations
Outcome alignment attestations prove that the final state matches the declared intent:
Target State: The desired end state declared in the intent.
Achieved State: The actual end state after execution.
Alignment Proof: ZK proof that the achieved state matches the target state within acceptable tolerances.
Tolerance Analysis: Detailed analysis of any differences between target and achieved states, with proof that differences are within acceptable tolerances.
Example:
Agent Behavior Attestations
Agent behavior attestations prove that agents followed agreed execution plans:
Execution Plan Commitment: Cryptographic commitment to the execution plan that the agent proposed in its bid.
Execution Proof: Proof that the agent's actual execution matched the committed plan.
Coordination Proof: In coalitions, proof that agents coordinated correctly according to agreed protocols.
No Deviation Proof: Proof that agents did not deviate from agreed plans without justification.
Auditability
VORT's proof system enables comprehensive auditing by third parties without requiring special access or trust in VORT infrastructure.
Independent Verification
Anyone can verify proofs independently:
Public Verification Keys: VORT publishes public verification keys that enable anyone to verify proofs without special access.
Open Verification Algorithms: Verification algorithms are published and can be implemented independently, ensuring no reliance on VORT's verification infrastructure.
On-Chain Cross-Verification: Proofs reference on-chain transactions, enabling cross-verification between proofs and on-chain state.
Third-Party Tools: Independent developers can build verification tools using published algorithms and keys.
Audit Workflows
Auditors can follow standardized workflows:
Proof Retrieval: Retrieve proofs for intents of interest through API, SDK, or on-chain interfaces.
Verification: Verify proofs using published verification keys and algorithms.
On-Chain Validation: Cross-reference proofs with on-chain state to validate consistency.
Constraint Checking: Verify that all constraints were respected by checking constraint satisfaction attestations.
Outcome Validation: Verify that outcomes match intents by checking outcome alignment attestations.
Report Generation: Generate audit reports documenting findings, verification results, and any issues discovered.
Continuous Auditing
The proof system enables continuous auditing:
Automated Monitoring: Automated systems can continuously monitor proofs, verifying them and alerting on any verification failures.
Anomaly Detection: Statistical analysis of proofs can detect anomalies that might indicate issues.
Trend Analysis: Analysis of proof patterns over time can identify trends, improvements, or degradations.
Compliance Verification: Proofs can be used to verify compliance with regulations, policies, or standards.
Audit Transparency
Audit processes are transparent:
Public Audit Reports: Auditors can publish audit reports that reference specific proofs, enabling independent verification of audit findings.
Open Audit Tools: Audit tools and methodologies are open source, enabling anyone to perform audits.
Audit Trail: Complete audit trails are maintained, documenting all verification activities and findings.
Replayability
VORT proofs enable replayability—the ability to replay intent execution using only on-chain state and cryptographic proofs, enabling independent verification of agent decisions.
Replay Mechanism
Replay works by reconstructing execution from proofs:
State Reconstruction: Starting from on-chain state at intent submission time, reconstruct the state as it existed during execution.
Action Replay: Replay each agent action using proof data, verifying that actions were valid and produced expected state changes.
Decision Verification: Verify that agent decisions (route selection, timing, etc.) were reasonable given available information.
Outcome Validation: Verify that the final outcome matches what the proof attests.
Replay Tools
VORT provides tools for replaying executions:
Replay API: API endpoint for triggering replays:
Replay SDK: SDK methods for programmatic replay:
Replay Dashboard: Visual replay interface in the dashboard showing step-by-step execution replay.
Use Cases
Replayability enables several important use cases:
Dispute Resolution: When disputes arise about execution correctness, replays can provide objective evidence:
Replay execution using only on-chain state and proofs
Verify that agent decisions were correct
Identify any deviations or errors
Provide objective resolution based on replay results
Debugging: When execution produces unexpected results, replays can help identify causes:
Replay execution step-by-step
Identify where execution diverged from expectations
Understand agent decision-making
Identify root causes of issues
Learning and Improvement: Agents can use replays to learn and improve:
Replay successful executions to understand what worked
Replay failed executions to understand what went wrong
Analyze decision patterns
Improve strategies based on replay insights
Verification: Independent parties can verify execution correctness:
Replay execution independently
Verify that replay results match claimed outcomes
Identify any discrepancies
Provide independent verification reports
Replay Guarantees
Replay provides strong guarantees:
Determinism: Replays are deterministic—given the same starting state and proof, replays always produce the same results.
Completeness: Replays can reconstruct complete execution, including all agent decisions and state changes.
Verifiability: Replay results can be independently verified by anyone with access to on-chain state and proofs.
Privacy Preservation: Replays verify correctness without revealing agent proprietary strategies or sensitive user data.
This proof verification system ensures that VORT maintains trustless operation while enabling comprehensive auditing, dispute resolution, and continuous improvement.
Last updated