Projects

Reusable engineering artifacts for AI reliability.

This page will collect open-source tools, reference implementations, documentation, and project notes connected to production AI agent infrastructure.

Published Projects

agent-trace-eval

Python library and CLI for golden trace regression testing of tool-using AI agents. Evaluates tool selection, arguments, ordering, handoffs, and recovery decisions against declarative YAML/JSON cases.

GitHub · PyPI · Related essay

Roadmap

Agent Trace Schema

A structured representation for agent runs, tool calls, retrieval events, policy checks, and evaluation outcomes.

Agent Trace Diff

A focused diff tool for comparing golden and candidate agent traces with readable failure reports.

Reliability Dashboard Template

A reference dashboard for monitoring agent success rates, failure modes, latency, and evaluation trends.

Reusable engineering artifacts for AI reliability.

Published Projects

agent-trace-eval

Roadmap

Agent Trace Schema

Agent Trace Diff

Reliability Dashboard Template