Configurable Automation Agent

A modular Python agent with persistent state, iterative planning, sandboxed tools, and automated program verification.

Overview

I built a configurable Python agent for code analysis, debugging, documentation, file inspection, and program verification. The system is designed around explicit tools and bounded execution rather than unrestricted access.

Planning and state

The agent maintains persistent state and can plan, act, inspect results, and self-correct across as many as 20 reasoning cycles. Its modular structure makes tools and task behavior easier to extend without rewriting the core loop.

Sandboxed execution

File and Python operations run inside a controlled environment. This reduces the risk of unintended changes while still allowing the agent to inspect programs and validate their behavior.

Verification

An automated validation layer checks program output and task results. The implementation achieved a 100% pass rate across nine verification tests.

What I learned

Useful agents need more than a capable model: they need clear state, narrow tools, observable outcomes, recovery paths, and tests that make success measurable.

Related projects

Forma - Open-source Form Builder (Full-stack Project)

A production-grade form builder with conversational forms, AI generation, payments, automation, and real-time integrations.

Next.js, TypeScript, Amazon Bedrock, Stripe Connect, Docker, Resend

VECINTI Marketplace & Agent Platform

A production local-services marketplace combining a timezone-aware booking engine, geospatial discovery, agentic AI, and Stripe Connect infrastructure.

Next.js, TypeScript, PostgreSQL, PostGIS, Amazon Bedrock, Stripe Connect