# From Signal to PR: Anatomy of a Self-Improving Agent — Jason Lopatecki, Arize

Source: [AI Engineer](https://www.youtube.com/watch?v=9HbzAWnKbo4)  
Feed7 permalink: https://feed7.dev/p/from-signal-to-pr-anatomy-of-a-self-improving-agent-jason-lopatecki-ariz-1i6f0ba  
Published: 2026-07-24T20:15:32.000Z  
Trust: Source Linked (source_linked)

## Why Included

Agents can turn production traces, evals, repository context, and observability skills into investigated issues or draft fixes. Human review remains more important as changes grow.

## Source Summary

Arize’s proposed loop starts from an event or periodic trigger, gathers traces, logs, eval results, and repository context through skills, then opens an issue or PR. The key input is evidence about the actual production code path, not a blind request to inspect the codebase.

## Practical Implication

Instrument agent systems so a coding agent can retrieve compact, relevant evidence inside its sandbox. Compose observability skills with repository access, choose the harness and sandbox explicitly, and make the reviewable artifact the output.

## Agent-Ready Context

Arize’s proposed loop starts from an event or periodic trigger, gathers traces, logs, eval results, and repository context through skills, then opens an issue or PR. The key input is evidence about the actual production code path, not a blind request to inspect the codebase.

Instrument agent systems so a coding agent can retrieve compact, relevant evidence inside its sandbox. Compose observability skills with repository access, choose the harness and sandbox explicitly, and make the reviewable artifact the output.

This is currently closer to a **responder-to-reviewer** shift than autonomous repair. Small fixes are the best fit; larger changes still need a human to steer them, and production access may require a **VPC-hosted sandbox** with carefully scoped connections.

## Context Map

- Layer: agent
- Domains: coding
- Topics: harness-engineering, observability, agent-reliability

## Uncertainty

- This is currently closer to a **responder-to-reviewer** shift than autonomous repair. Small fixes are the best fit; larger changes still need a human to steer them, and production access may require a **VPC-hosted sandbox** with carefully scoped connections.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.