Last updated 8 months ago
Billions of dollars have been lost on blockchains due to smart contract exploits, and design flaws. In Q2 2025 alone, over $800M was stolen from DeFi protocols through vulnerabilities (Certik).
We will develop an AI-powered smart contract auditing and optimization platform for Cardano. The tool will detect vulnerabilities in Plutus contracts using AI/ML models trained on real exploit data.
This is the total amount allocated to AI-Powered Smart Contract Auditing Tool for Cardano.
Please provide your proposal title
AI-Powered Smart Contract Auditing Tool for Cardano
Enter the amount of funding you are requesting in ADA
70000
Please specify how many months you expect your project to last
7
Please indicate if your proposal has been auto-translated
No
Original Language
en
What is the problem you want to solve?
Billions of dollars have been lost on blockchains due to smart contract exploits, and design flaws. In Q2 2025 alone, over $800M was stolen from DeFi protocols through vulnerabilities (Certik).
Supporting links
Does your project have any dependencies on other organizations, technical or otherwise?
Yes
Describe any dependencies or write 'No dependencies'
AI Models that are going to be trained for smart contract auditing.
Will your project's outputs be fully open source?
Yes
License and Additional Information
The solution will be open-sourced (MIT License), hosted on GitHub, and built with community engagement to ensure transparency and adaptability.
Please choose the most relevant theme and tag related to the outcomes of your proposal
Security
Mention your open source license and describe your open source license rationale.
The solution will be open-sourced (MIT License), hosted on GitHub, and built with community engagement to ensure transparency and adaptability. Anyone who wants to contribute to the project can create an issue and pull request on GitHub.
How do you make sure your source code is accessible to the public from project start, and people are informed?
We will make the souce code available for everyone by hosting it on GitHub. Beside that we will use different social media mechanism to inform everyone the progress and features the tool is building or finished integrating.
How will you provide high quality documentation?
We will provide a high quality documentation by following the standard documentation writing methods. And we will also iterate and update on the documentation if the features included on the tool are not already on the documentation.
Please describe your proposed solution and how it addresses the problem
Our AI-powered auditing platform will combine static analysis, dynamic execution tracing, and AI reasoning frameworks to deliver high-accuracy vulnerability detection and performance optimization. Inspired by recent research such as SmartAuditFlow (2025), our approach leverages a Plan-Execute AI framework with Retrieval-Augmented Generation (RAG) to minimize hallucinations and false positives.
Core Features:
Please define the positive impact your project will have on the wider Cardano community
What is your capability to deliver your project with high levels of trust and accountability? How do you intend to validate if your approach is feasible?
We will also collaborate with Cardano developer groups and security researchers to validate results and improve effectiveness.
Milestone Title
Architecture & research
Milestone Outputs
We will begin by designing the system architecture and conducting deep research on model design and dataset preparation. This includes collecting high-quality datasets of smart contracts, both secure and vulnerable, and curating them for AI model training. At this stage, we will also prepare detailed documentation of the overall approach, establish coding standards, and set up the GitHub repository for public collaboration. This milestone ensures a strong foundation for all later stages
Acceptance Criteria
This milestone will be accepted when a complete system design document is prepared, datasets are curated and structured in a usable form for training, and a GitHub repository is publicly available with initial commits, readme files, and contribution guidelines that allow others to review and participate.
Evidence of Completion
The evidence will include finalized architecture documentation, structured datasets in a repository or dataset store, and a publicly available GitHub repository showing commits, initial project structure, readme files, and contribution instructions.
Delivery Month
1
Cost
5000
Progress
10 %
Milestone Title
AI model development
Milestone Outputs
This phase focuses on creating and training the AI/ML model to detect vulnerabilities in Plutus smart contracts. The model will be trained using real exploit datasets and simulated vulnerabilities, allowing it to identify flaws such as reentrancy, integer overflows, and access control issues. We will also integrate it with test contracts to validate accuracy and usability.
Acceptance Criteria
The milestone will be accepted once the trained AI model demonstrates accurate detection of vulnerabilities in Plutus contracts during testing, successfully flags a range of known flaws, and integrates with a test environment that developers can run sample audits on without major errors.
Evidence of Completion
Evidence will consist of test reports highlighting the model’s detection rates, performance metrics such as precision and recall, and GitHub commits with the trained model, test scripts, and integration documentation proving that the system runs as intended.
Delivery Month
2
Cost
30000
Progress
40 %
Milestone Title
A web interface
Milestone Outputs
We will develop a web-based interface and a CLI tool to make the auditor accessible. These tools will allow developers to upload or input contracts and receive immediate audit feedback. We will also integrate an optimization engine that recommends improvements to reduce transaction costs and strengthen contract security.
Acceptance Criteria
This milestone will be accepted once the web interface and CLI tool are operational, allowing users to upload or input Plutus contracts and receive reliable vulnerability detection results alongside optimization suggestions. The system must function without major errors and be easy to use.
Evidence of Completion
Evidence will include a deployed and accessible website, a working CLI package with clear installation instructions, and test cases or demo videos showing successful contract analysis with both vulnerabilities flagged and optimizations recommended.
Delivery Month
1
Cost
10000
Progress
20 %
Milestone Title
Refine the Model using users feedback
Milestone Outputs
We will conduct pilot audits with 2–3 real Cardano projects and gather developer feedback through surveys, workshops, and hands-on sessions. This feedback will guide the refinement of the AI models and improvements to documentation. The aim is to align the tool with actual developer needs and ensure community-driven improvements.
Acceptance Criteria
This milestone will be accepted when developer feedback has been formally collected and documented from at least two pilot projects, and the AI model has been refined to address the identified shortcomings. Updated documentation must also reflect these changes to ensure transparency.
Evidence of Completion
Evidence will consist of workshop and survey reports, feedback summaries, updated AI model versions with changelogs describing refinements, and improved documentation that reflects developer input and adjustments made.
Delivery Month
2
Cost
15000
Progress
20 %
Milestone Title
Public launch
Milestone Outputs
The final milestone will focus on publicly releasing the tool under the MIT License. This includes publishing final documentation, audit case studies, and pilot project reports. We will also organize outreach through developer communities and Cardano ecosystem channels to drive adoption and encourage contributions.
Acceptance Criteria
This milestone will be accepted once the tool is officially released on GitHub with the MIT License, comprehensive documentation is publicly available, final case studies from pilot projects are published, and announcements are made through Cardano developer channels to confirm launch.
Evidence of Completion
Evidence will include the public GitHub repository with complete code and license, final case studies and reports in PDF or online format, official launch announcements across Cardano forums, and fully accessible documentation for developers.
Delivery Month
1
Cost
10000
Progress
10 %
Please provide a cost breakdown of the proposed work and resources
AI Model Development & Training → ₳25,000
Smart Contract Security Research → ₳10,000
Frontend & Backend Development → ₳15,000
Pilot Testing & Case Studies → ₳10,000
Documentation, Community Outreach, Workshops → ₳10,000
Total = ₳70,000
How does the cost of the project represent value for the Cardano ecosystem?
The requested ₳70,000 provides exceptional value to the Cardano ecosystem because it delivers a critical piece of infrastructure: an open-source AI-powered smart contract auditing and optimization tool that can be used by all developers, present and future, at no cost.
Why this represents strong value:
Terms and Conditions:
Yes
Natnael Mesfin: Frontend/Platform Develope
Fraol Bereket : A fullstack blockchain developer and ML specialisty.
GitHub: https://github.com/fraolb