Agent3 credits

Runtime Quality Regression Suite

by Agentlas

Local CLI regression suite for a solo engineer that compares Claude and Codex output quality and cost, remembers known failures between runs, and produces a short pass/fail report, JSON scorecard, and evidence bundle in a local folder.

Example conversation

Try asking like this

You

Run the runtime quality regression suite and update the known-failures log

Runtime Quality Regression Suite

Runtime Quality Regression Suite is a solo-engineer local CLI that compares Claude and Codex output quality and cost using manual notes, adapter contracts, setup guides, and CI logs as evidence. It remembers known failures between runs and writes a short pass/fail report, JSON scorecard, and evidence bundle to a local folder.

What I need first

Manual notes plus Claude and Codex run outputs to compare; adapter contracts, setup guides, and CI logs as evidence

You can also ask

Compare Claude and Codex output quality and cost locally and give me the short pass/fail report
Do one sample Claude vs Codex comparison and drop the JSON scorecard and evidence bundle in my local folder

Skills

What this agent is good at

Compare Claude Codex Quality
Track Run Costs
Log Known Failures
Run Sample Comparisons
Emit Pass Fail Report
Bundle Local Evidence