Independent builder focused on governed AI agents, Web3 infrastructure, and verifiable execution.
I build open-source tools and standalone hackathon projects that make consequential agent actions easier to inspect, approve, and audit.
| Project | What it does | Context |
|---|---|---|
| IntentProof (demo) | A local-first Solana transaction policy firewall that evaluates transfers before signing and records hash-linked evidence after user confirmation. | Superteam Agentic Engineering Grant |
| SolePilot (live app) | A governed agent runtime for solo operators, with deterministic policy gates around tasks, messaging, and Solana payments. | OpenArena BUIDL_QUESTS 2026 |
| TxSentinel (live app) | A non-custodial, pre-sign transaction firewall that returns ALLOW, HOLD, or DENY before an autonomous agent can proceed. |
OKX.AI Genesis Hackathon |
| Meme Scanner | A local Solana meme-token risk scanner with explicit safety gates, explainable outcomes, and a deterministic replay mode. | Independent research project |
- Audit Skill Supply Chain - a pre-install security gate for open-source agent skills, with quarantine, provenance checks, and integrity verification.
- Solana Policy Firewall - an MIT-licensed, fail-closed pre-sign policy engine written in Rust.
- AgentPay Firewall - an x402 payment-rail reference implementation for deterministic agent spending controls.
TypeScript, Python, Rust, Solana, EVM, x402, GitHub Actions, and Docker.
