AuditAgent

AuditAgent

AuditAgent was conceived in response to the growing demand for secure, transparent smart contract deployments in the blockchain industry. It helps both developers and auditors automate routine checks, surface early issues, and prepare code for deeper manual review by streamlining repetitive tasks and highlighting hidden bugs.

Client

© 2025 Nethermind. Displayed with permission. All rights reserved.

Industry

AI Code Auditor - Web3 & SaaS

Services

UI/UX Design, UX Research, Interaction Design, Prototyping, and Product Strategy.

Year & Role

2025, Lead UX & Product Designer

Client Recommendation

Roy Neubert,
Design Head at Nethermind

Reetesh is an outstanding team member—super reliable, collaborative, and always delivers timely, well-considered designs. A joy to work with, he plans meticulously even for complex tasks and is always ready to help.

Summarized by AI

Tech Stack

Problem Discovery

Traditional audits are often lengthy, costly, and out of reach for small Web3 teams, leaving security gaps and launch delays.

Full security audits often takes 2-5 weeks, slowing down both development and launches

Full security audits often takes 2-5 weeks, slowing down both development and launches

Manual review of common issues increases workload for auditors and developers

Manual review of common issues increases workload for auditors and developers

Teams face inconsistent reporting and unclear next steps before formal audits.

Teams face inconsistent reporting and unclear next steps before formal audits.

Automated tools often produce too many false positives, leading to wasted time and low trust.

Automated tools often produce too many false positives, leading to wasted time and low trust.

Design Process

Each design step was set up to deeply understand users and quickly adapt to feedback. This let us align our solutions with real workflow.

1

Research & Discovery

We researched existing tools to learn what worked well, and what to avoid, while shaping a solution that felt uniquely ours.

2

Design & Prototype

Early designs & prototypes helped capture the core idea and align the team before going deeper into design decisions.

3

Beta Phase with Early Users

By testing with early adopters, we uncovered blind spots and usability issues that weren’t obvious during internal reviews.

4

Iterative Improvements

Based on real usage and feedback, we refined the interface, adjusted features, and improved clarity—focusing on what users needed the most.

Market Potential

Blockchain smart-contract security is a small market but growing rapidly: valued at $180.53 million in 2023 and forecast to reach $839.37 million by 2030, with a 24.55% CAGR. Additionally, $2.1 billion was lost to hacks in 2024, underscoring the demand for faster and more affordable protection.

User Insights

User research surfaced a need for instant scans, clear guidance, and trusted automation. These insights below directly shaped our product decisions.

Devs want instant checks; waiting days slows progress

Users need rapid, clear explanations of risks and fixes

Teams prefer a shared dashboard to streamline collaboration

Continuous scanning is more valuable than occasional audits

Automation should simplify workflows, not add extra noise

Business Goals

We set measurable business targets to stay focused on impact, helping every design choice build toward real outcomes.

Improve access to affordable pre-audits for dev adoption

Improve access to affordable pre-audits for dev adoption

Reduce vulnerability-fix time with clear recommendations

Reduce vulnerability-fix time with clear recommendations

Build trust in AI-based pre-audit tools, through transparency

Build trust in AI-based pre-audit tools, through transparency

Capture market share in AI-powered auditing tools for Web3

Capture market share in AI-powered auditing tools for Web3

User Persona

Based on user interview insights, we crafted two personas to represent our target user and their core pain points.

Sophia Williams

San Francisco

25 Age

Blockchain Developer

She’s technical, uses CI/CD, and depends on tools to catch bugs early and keep smart contracts clean.

Pain Points

Not knowing if her code is ready for a formal audit.

Unclear outputs often require back-and-forth interactions with auditors.

Difficult to interpret noisy automated scan reports.

Goals & Needs

Quickly check contracts for issues before submitting for manual audit.

Clear, actionable explanations for each issue, not just code snippets.

Integrate findings into her team’s GitHub and task tracker.

Make her codebase “audit ready,” not to skip manual review.

James Brown

Quebec

27 Age

Smart Contract Auditor

He audits DeFi and DAOs. He wants automation to handle routine issues so he can focus on critical risks.

Pain Points

Manual, repetitive review of basic code hygiene.

Risk of missing easy-to-overlook bugs, especially under time pressure.

Ensuring audit quality and thoroughness for team.

Goals & Needs

Reduce repetitive checks (e.g. basic linting, code-style errors, common issues).

Get a summarized risk overview before manual engagement.

Use AI findings as an extra safety net to catch hard-to-spot or overlooked bugs.

Seamlessly document findings for clients and compliance.

User Journey Map

Mapping the old auditing process exposed user pain points and guided us to areas where AuditAgent could add the most value.

Click to expand

Product Market Fit

We positioned AuditAgent as the bridge between quick static scans and deep manual audits—offering affordable, fast checks that catch business logic issues and prep code for full review.

Click to expand

Feature Prioritization

We prioritized features based on urgency and user value to keep the MVP focused, lean, and impactful.

Click to expand

Solution & Vision

We should address blockers in mainstream smart contract security and build an AI auditing tool reliable, friendly, and integrable.

AI pre-auditing tool with latest LLMs and attacker reasoning

AI pre-auditing tool with latest LLMs and attacker reasoning

Clear, actionable findings with severity, fixes, and snippets

Clear, actionable findings with severity, fixes, and snippets

Flexible pay-per-scan pricing, with free scans for small audits

Flexible pay-per-scan pricing, with free scans for small audits

Easy adoption via developer-friendly CI/CD integration

Easy adoption via developer-friendly CI/CD integration

Information Architecture

AuditAgent’s structure is organized and user-focused, grouping features intuitively for easy navigation.

Click to expand

Design Principles

Clear principles kept our UX unified and purpose-driven, helping the team move fast without losing focus.

Clarity over Complexity

We chose clean layouts and plain language so users could understand issues without deep security knowledge.

Clarity over Complexity

We chose clean layouts and plain language so users could understand issues without deep security knowledge.

Action-First UX

Each result gave a clear next step—suggested fixes or copyable code—to speed up resolution and action.

Action-First UX

Each result gave a clear next step—suggested fixes or copyable code—to speed up resolution and action.

Built for Dev Flow

The experience fits into existing developer habits, with GitHub integration and CI/CD support.

Built for Dev Flow

The experience fits into existing developer habits, with GitHub integration and CI/CD support.

Branding and Component Library

I designed the brand identity and used HeroUI components to ensure a consistent look and seamless developer handoff.

Dashboard

We crafted a unified dashboard so users could view scans, reports, and team activity at a glance. This simplified navigation and helped users quickly track their audit progress.

Scan Code Flow

Our scan flow was made direct and quick, guiding users through repo connection and scan setup in under a few minutes. Clear guided texts and progress indicators reduced drop-off and stress.

Audit Results Flow

Audit results spotlight key charts and stats up top, so users get instant insights. We prioritized what matters most—organized vulnerabilities with clear explanations, and a handy table of contents.

Mobile Responsive Designs

Key workflows and critical screens were optimized for mobile, so users could run and review scans easily wherever they are.

Website Design

I designed and launched the public website using Framer, allowing for quick updates without relying on a developer.

Design Iterations

We improved the product through user feedback—refining the scan setup, audit results flow, and adding flexible pricing. Left shows earlier versions and the right side shows the final improvements.

Results and Impact

These outcomes reflect how well our solutions performed, based on actual usage data from the MVP and beta release.

18K+ vulnerabilities discovered, providing accessible security coverage to developers globally.

18K+ vulnerabilities discovered, providing accessible security coverage to developers globally.

4–6 hours saved per 1,000 LOC, reducing pre-audit timelines from weeks to hours.

4–6 hours saved per 1,000 LOC, reducing pre-audit timelines from weeks to hours.

3× faster code reviews by helping teams quickly prioritize and assign issues.

3× faster code reviews by helping teams quickly prioritize and assign issues.

AuditAgent caught logic bugs missed by other tools, solidifying its role as a trusted AI auditor.

AuditAgent caught logic bugs missed by other tools, solidifying its role as a trusted AI auditor.

Learnings

These learnings reflect my growth as a designer while working on AuditAgent, from understanding developer workflows to iterating on real feedback, and what I’d carry into future projects.

I was new to designing for dev tools, especially in web3, so I had to learn how smart contract developers think and work.

We realized developers cared more about scan quality and fixability than flashy visuals.

Having smart contract auditors on our team gave us quick feedback and kept designs grounded in real audit workflows.

Some edge cases only appeared after launch, so we kept refining flows even after the MVP shipped.

A lot of my time went into simplifying things users didn’t even realize were confusing.

Work With Me

Got an idea for an interesting product or need tips on product design? Don’t hesitate to get in touch by filling out the form below or reaching out to me via email reeteshuiux@gmail.com

Avg. response 12 hours