# Lessons from my self-improving coding agent

DevFeed: [Lessons from my self-improving coding agent](<https://devfeed.tech/articles/lessons-from-my-self-improving-coding-agent-83198.md>)

Original publisher: [Read original article](<https://ngrok.com/blog/bmo-self-improving-coding-agent>)

Author: Joel Hans

Published: 2026-02-25T00:00:00Z

Content type: article

Language: en

Sources: [\[ngrok news\]](<https://devfeed.tech/sources/ngrok-news.md>)

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [agentic-coding](<https://devfeed.tech/topics/agentic-coding.md>), [agent orchestration](<https://devfeed.tech/topics/agent-orchestration.md>)

Tags: [agent-harness](<https://devfeed.tech/tags/agent-harness.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-coding](<https://devfeed.tech/tags/agentic-coding.md>), [agentic-loop](<https://devfeed.tech/tags/agentic-loop.md>), [ai](<https://devfeed.tech/tags/ai.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>)

## AI overview

The author describes building and using bmo, a coding agent designed to improve its tools and workflows through active learning, session reflection, and telemetry. In nearly 100 sessions, specialized tools helped make common tasks more reliable, but the agent often deferred improvements and could not independently sustain self-diagnosis. The author concludes that its gains come from prompts, tools, and scaffolding rather than changes to the underlying model, with human guidance remaining central.

## Source excerpt

Two weeks with a self-improving coding agent taught me more about LLMs than a year of prompting ever did.