Kleros Development Update: September 2025

September brought significant progress across the Kleros ecosystem. From completing major migrations to launching security competitions and implementing new features, the team continued building toward the next generation of decentralized justice. Here's what happened this month.

Kleros Core

Court V2

Frontend Development

The Court V2 interface received important enhancements this month. Markdown Editor and Markdown Renderer support was added for juror justifications, evidence submissions, and dispute descriptions, giving users more formatting flexibility. This feature has been merged and is now available.

The DK (Dispute Kit) features selection progressed through multiple iterations and review phases. The University Court was successfully redeployed.

The Enterprise UI (Rulings UI) implementation was completed and moved into review, bringing organizations closer to a streamlined arbitration interface.

A critical bug affecting vote reveals was identified and fixed. Users were encountering out-of-range number errors, but comprehensive test cases were created, the root cause identified, and a fix submitted to properly handle invalid disputes.

Contract Development

September marked a crucial milestone as V2 contracts entered internal review. The contracts were frozen to ensure stability during this process, with multiple reviewers examining the codebase thoroughly.

Several important features were merged:

• Dispute kit jump to non-classic implementation
• PNK penalties applied to sortition tree
• Shutter commitment recovery
• Sortition trees library improvements
• Storage layout and events ABI updated

Additional work in progress includes:

• Fuzzing tests (in review phase)
• Foundry tests were split into multiple files for better organization

The team also conducted analysis for the Argentina Ombudsman use case and held discussions around dispute kit jumping logic optimizations.

Deployment & Infrastructure

Deployment stability was a focus this month. After contracts temporarily became incompatible with supporting infrastructure due to breaking changes ahead of internal reviews, the team worked to stabilize the system. Neo preview branches were restored, and University contracts were redeployed on devnet.

Reaching 100 cases on the v2 beta was celebrated as an important milestone, with discussions around next steps and audit preparation.

Court V1

Wallet support improvements progressed through development and are now in the merging phase, enhancing user experience for jurors.

🌐 Experience next-generation decentralized justice! Court V2 approaches full production deployment.

Curate

Platform Updates

The Curate ecosystem focused heavily on indexer improvements. Subgraph development for migrating from TheGraph to Envio was completed, though final checks continue before switching endpoints across all applications. Issues with TheGraph's item count were resolved, and Envio resyncing and testing progressed throughout the month.

Testing with the new indexer for Curate v1 was completed successfully. The pull request is scheduled to merge in early October, which will enable submissions on Scout ATQ to resume.

Scout V2

Scout V2 received substantial user experience enhancements:

CSV Export Feature: A highly requested feature was implemented allowing users to export entire lists to Excel files. The export is customizable by status, network, and timeframe, making it valuable for data analysis and simplifying workflows for submitters and challengers.

Improved Filtering: The filter and export modals received UX improvements. Users can now hover over any status or network to see an "Only" button that instantly isolates that option. An "All" button allows quick selection of everything again.

Design & Polish: New layout designs are in progress to further enhance the website. The Dashboard Home page was refined, filters were added to the Activity page, bugs in filter modals were fixed, and mobile styling was optimized. Carousel bugs in the dashboard were also addressed.

The team actively collected and prioritized feedback from the community to guide ongoing development.

Scout V1

An additional entry validation check was introduced: when new submissions are initiated on EVM chains, the system now verifies proper formatting (0x...) to prevent submission errors. Migration to a more reliable indexer continues in the testing phase.

Scout Snap

Domain display during token transfers was updated, and contract interaction details were improved. General copywriting was enhanced for better user friendliness.

Compatibility work began with MetaMask's team to enable Scout Snap for EIP-7702 compatible Smart Accounts. A planned update to a more reliable indexer is in the pipeline, following the approach of other Curate applications.

Stake Curate

Parameters and policy are being finalized for the first list launching in partnership with DAMM Capital. This inaugural list will establish patterns for additional lists to launch on Stake Curate in the coming months.

🕵️‍♂️ Join the curation revolution! Explore Scout V2 and contribute to Stake Curate on Gnosis Chain.

Proof of Humanity V2

zkPOH Implementation

A significant milestone was reached with the first draft of the zkPOH contract using the Semaphore protocol. The implementation includes:

• Complete test suite with 100% coverage
• Deployment scripts for easy setup
• CLI utility to generate proofs and commitments

This enables privacy-preserving humanity verification, a major step forward for the protocol.

Keeper Bot

A critical bug was discovered where the keeper bot published parallel transactions with the same nonce, causing all but the first to fail. The bot was completely refactored to process transactions sequentially.

The refactor brought substantial improvements:

• Composable services following clean code principles
• Optimized subgraph queries
• 100% unit test coverage
• Robust custom error handling

User Experience

The Request Page received multiple enhancements:

• Improved UX for Execute/Advance Actions and Funding Modal
• Loading indicators and disabled states during pending transactions
• Auto-refresh and clearer user feedback mechanisms

✅ Verify your humanity! You can check our current implementation here

VEA (Vea Bridging)

VEA focused on robustness and reliability this month:

• Failed resolution scenario fix in the validator bot was completed, handling edge cases where Inbox snapshots mismatch Outbox claims
• Secondary validator bot implementation began to provide additional redundancy
• Contract test coverage was expanded
• Contracts were redeployed (publication pending)
• Hashi Lightbulb POC was completed

The team evaluated the impact of Hashi's discontinuation from Gnosis DAO and planned next steps, including adding the Hashi executor to the relayer-cli.

🌉 Bridge with confidence! Experience enhanced cross-chain messaging with Vea.

Atlas

Production Migration: The Atlas migration to production was successfully completed this month, representing a major infrastructure upgrade. The new macroservices architecture was deployed with database migrations performed with minimal downtime. The backend team delivered this critical update, improving performance and reliability across the platform.

The monitoring system was also migrated to support the multi-services architecture, with ongoing improvements to monitoring capabilities.

New Features & Endpoints: Court V1 notifications are almost ready and currently in testing. The next step involves coordinating with the frontend team to migrate the Court site to use this notification system. The users database also needs to be migrated from the old system to Atlas, with this work planned for October if no major issues emerge during testing.

New endpoints for Court V2 frontend were created to retrieve blockchain data, including voting history and stakeSet history. These endpoints are undergoing testing and should be released during October.

Bot Improvements: Refactoring and bug fixing were performed to address issues in the PoH V2 Keeper Bot, with testing currently underway in the staging environment before moving to production. Part of the team is also working on Zodiac bots.

Infrastructure Maintenance: Regular maintenance work continued on the forums (Kleros forum this month) and other infrastructure components.

Future Plans For the coming month, the team plans to begin migrating old bots that are difficult to maintain to Atlas, including Court V1 keeper bots and Realitio Cross Chain bots.

Reality

Security Focus

A major security initiative launched this month with the Hats Competition to identify vulnerabilities in the Cross-Chain Reality Proxies. After final review and vault deployment, the competition officially started on September 23rd.

SafeSnap security is being enhanced with bot updates to perform automated security checks, improving the safety of governance actions.

Deployments & Updates

SUKU deployment was completed, and work began on Zodiac SafeSnap automation. Cross-chain reality proxy scripts were updated to match the latest contracts package changes, and updated proxy contracts were deployed.

Multiple issues were identified and resolved throughout the month:

• Handling failed dispute creations
• Underflow issues
• Smart wallet compatibility
• Metaevidence retrieval problems

The team reviewed bug reports and vulnerability claims from the security competition, discussing validity and severity, and worked toward closing resolved bug bounties.

Futarchy

The Futarchy prediction market platform received several improvements:

• Current position value display was enhanced
• Button added to claim prizes from parent markets with unspent outcome tokens
• Claim flow for prediction tokens was improved
• Bugs related to complex prediction routes were fixed
• Chart optimizations and new Chart UI implemented
• "Fill to price order" functionality added
• Distilled Judgement experiment branch deployed
• Transaction batching planning began

Escrow

Minor fix was made to evidence display (for scenarios where evidence was uploaded via the Dispute Resolver), significant work happened behind the scenes.

The new Escrow v1 frontend development is essentially complete and will soon move to testing. The rewrite uses the components library for consistency with other Kleros products. Disputes and appeals support implementation is nearly done.

The SmartInvoice escrow review was also completed this month.

🔒 Secure your transactions! You can try our Escrow V2 and Escrow V1

Other Products

Dispute Resolver v1: Several fixes are in progress to address reported issues.

Mirrorfall.ai: Work continued on dispute creation on v2 testnet, a sample report was produced, and a new domain was secured.

KIP-78: The staking reward proposal was reviewed and subsequently implemented.

Infrastructure & Tooling

The development team continued improving internal processes and monitoring:

• Resolved various deployment and configuration issues, including Netlify endpoint configurations and contract-subgraph compatibility challenges
• Exploration of multiple approaches for subgraph monitoring, including using Atlas
• Consideration of RPC reliability solutions for Envio, including multiple RPCs with quorum systems
• Introduction of tools like VSDeer for scanning and verifying IDE extensions
• Court V2 documentation updates with ongoing team reviews
• PNK faucet refills on Arbitrum Sepolia testnet

What's Next

As October begins, key priorities include:

• Completing Court V2 internal reviews and preparing for external audits
• Finalizing Envio migration across Curate applications
• Continuing the Hats security competition and addressing findings
• Advancing zkPOH development
• Launching the first Stake Curate list
• Testing and releasing the new Escrow v1 frontend

September saw the completion of critical migrations, strengthening of security practices, and continuous refinement of user experiences. The momentum builds toward significant releases ahead.

Thanks for following along with our progress!

Get Involved

Want to contribute to Kleros? Join our Discord, follow us on Twitter/X, and explore opportunities on kleros.io.