Run coding agents like a medical team to fix bugs

Community report: cockroach Labs assembled a GitHub Actions pipeline in which labels on an issue trigger role-specific agents. A Fellow tries to reproduce the bug and posts a treatment plan, and a separate Review Attending approves or rejects it before any code is written. Over five months it merged 1,238 PRs and over a million lines of code with 7 reverts.

Agent or tool: Custom agent
Topic: Software Development
Reported on: Hacker News
Added: 2026-10-15

Why it matters

Coding agents are moving from autocomplete to doing whole tasks: reading a codebase, making a plan and changing many files. This case is a concrete example of where that works in practice. It is one of 106 sightings filed under Software Development on Agent Sightings (number 63 by interest), and one of 182 involving Custom agent.

How you could try it

Here is a reusable prompt, written by us in our own words, that captures the task. Paste it into your own agent or any similar agent and adjust the details to your situation.

In this repo, run coding agents like a medical team to fix bugs. Read relevant code, apply a minimal patch, run available checks, and summarize the diff. Stop before merge or deploy.

The prompt breaks down into 3 steps you can check as the agent works:

In this repo, run coding agents like a medical team to fix bugs.
Read relevant code, apply a minimal patch, run available checks, and summarize the diff.
Stop before merge or deploy.

Tip: Work on a fresh branch, ask for a plan before edits, and review the diff and test results before merging.

Source

Read the original first-hand account on Hacker News: https://news.ycombinator.com/item?id=50021899. Our summary is a paraphrase; details are in the source.

Related sightings

Test how well coding agents can onboard to your product · Custom agent · Software Development
Case study takeaway: you enter your product's site on AX Check, which first checks the site itself. Three coding agents then try to onboard end to…
Turn one goal into an ordered plan of coding-agent tasks · Custom agent · Software Development
According to the write-up, a planner, which can be a coding agent running read-only, researches the repo and asks clarifying questions. It then…
Run a personal assistant that handles repo issues and groceries · Custom agent · Software Development
As described in the source, custom agent shows up in a write-up about run a personal assistant that handles repo issues and groceries. The linked…
Fix app bugs, harden security and ship fixes autonomously · Hermes Agent · Software Development
Omar’s health app got a Hermes babysitter during work hours: triage priority bugs, fence personal browser logins away from bots, sketch a tax-refund…
Rebook a work flight to save money · Custom agent · Travel & Booking
A published walkthrough shows arden monitored Someone's booked Zurich work flight. It sent a text proposing a rebook that would save $7,750 and…
Get a billing dispute refunded by a support agent · Custom agent · Finance
Case study takeaway: the company's AI agent produced and issued a refund to a customer who had a dispute with an incorrect Vercel bill.

← All 991 use cases