← All programs
AIA-301AdvancedAI40h · 5 weeks (self-paced)
AI-Augmented Smart Contract Auditing
A practical program for auditors who want to 2-3x their review speed with AI tooling. Covers prompt engineering for Solidity review, retrieval-augmented audit pipelines, integrating LLMs into Foundry and Slither workflows, and — critically — understanding where models hallucinate and how to catch it before a finding ships.
01 · Outcomes
What you will be able to do.
- 01Design structured prompts that extract reliable Solidity findings from LLMs
- 02Build a retrieval-augmented pipeline that feeds a codebase to a model and returns structured output
- 03Integrate AI assistants into Foundry and Slither analysis workflows
- 04Identify and flag hallucinated findings before they reach a report
- 05Evaluate new AI-audit tools rapidly and assess their fit for an engagement
02 · Syllabus
What you will cover.
Week 1
Foundations of LLM-Assisted Code Review
- →How transformer models process Solidity — strengths and blind spots
- →Prompt engineering patterns for security review (few-shot, chain-of-thought)
- →Setting up a reproducible AI-audit workbench
Week 2
Retrieval-Augmented Audit Pipelines
- →Chunking, embedding, and indexing large contract codebases
- →Feeding context windows with the right files at the right time
- →Structured output: getting findings as severity/location/remediation JSON
Week 3
Toolchain Integration
- →Piping Slither and Foundry output into an LLM for triage
- →Building custom MCP tools for audit-specific queries
- →Automating first-pass review with CI-triggered model runs
Week 4
Hallucination Detection & Quality Control
- →Common hallucination patterns in code-review models
- →Cross-checking model output against on-chain state and test results
- →Building a human-in-the-loop review gate
Week 5
Capstone & Tool Evaluation
- →Full AI-augmented audit of a non-trivial protocol
- →Speed and accuracy comparison: manual vs AI-assisted
- →Evaluating emerging AI-audit products with a structured rubric
03 · Lessons
Course content.
5 lessons available. Preview the material before enrolling.
Tools
- • Claude / GPT-4
- • Foundry
- • Slither
- • LangChain
- • Anthropic MCP
- • ChromaDB
Who it's for
- Working auditors looking to increase throughput with AI
- Security engineers building internal AI-assisted review pipelines
- Protocol teams that want to pre-screen code before a formal audit
Prerequisites
- Completion of SCA-201 or active smart-contract audit experience
- Basic comfort with Python scripting and API calls
- Familiarity with at least one LLM provider (Claude, GPT, etc.)
Ready?