His team makes coding agents check their own work before a human sees it
His team puts a few lines in their agent instruction files telling agents to run simple self-verification steps. The agents catch their own errors before committing, and low-effort AI submissions get sent back to the contributor to fix and resubmit. He says it filters out most submissions nobody bothered to look at.
How it works
- Team adds self-verification steps to the agent instruction files
- Agents check their own work before committing
- Weak submissions are returned to the contributor
Why it’s worth knowing
A couple of lines in an instruction file became a quality gate that bounces careless submissions before any human looks.
Where it’s from
Summarized from a public post on njump.me. See the original ↗