OpenAI Build Week project
Smart contract audits run like an agent.
Submit a repo and AuditPilot detects the stack, maps the attack surface, checks vulnerability patterns, and returns an auditor-style report for Solidity and Solana programs.
2launch ecosystems
5severity levels
1shared agent engine
Sign in and add your OpenAI key first. Web audits run on your own quota; MCP audits reason inside the connected agent's own session.
IngestionRepo detector ready
ReasoningUser-key GPT loop
OutputPersistent report
OpenAI key
RequiredAuditPilot uses your key for web audits. The key is encrypted server-side and never returned to the browser.
Sign in with GitHub before saving a key.
Audit Report
Run an audit to see severity-ranked findings, exploit reasoning, and fix recommendations.
Agent Timeline
Authenticate with GitHub through Supabase
Load your encrypted OpenAI key server-side
Create a persistent audit run before analysis starts
Update audit history when findings are ready