Building AI agents, agent security tools, and onchain apps.
- AI agents that take real actions safely
- Agent security: prompt injection, least privilege, approvals
- MCP servers and tool integrations
- Onchain apps (Solidity, Monad)
- Voice AI products
Broski - Accounts-payable agent that allows, holds, or blocks MCP tool calls
- Blocked a prompt-injected $50k wire in testing
- Built for ArmorIQ's Flow State hackathon
KidCheck - AI agent that helps students find opportunities and check eligibility
CallDesk - Voice front desk for businesses
- Built for the AssemblyAI Voice Hackathon
- Live: https://call-desk-mu.vercel.app
technocore-guideguard - Audits guides and scripts for unsafe key handling
Byuggy - Submitted to Seer's Foreguild contest
TypeScript, Next.js, Solidity, Foundry, Rust, MCP, Vercel
X: @chiboy0123