# A minimal eval harness you can run in CI

Source: [Vercel](https://vercel.com/blog/minimal-eval-harness-ci)  
Feed7 permalink: https://feed7.dev/p/eval-harness-post  
Published: 2026-06-26T00:00:00.000Z  
Trust: Source Linked (source_linked)

## Why Included

Shows how to gate agent changes behind a tiny, fast eval set in CI.

## Source Summary

A 20-case eval wired into CI that fails the build on regression. Full code in the post; runs in under 30 seconds.

## Practical Implication

Makes agent reliability a CI concern, not a vibe. Start with your 5 most common failure cases.

## Agent-Ready Context

Wire a 20-case eval into CI; fail the build on regression. Keep it fast (<30s) so agents get feedback each PR.

## Context Map

- Layer: benchmark
- Domains: coding
- Topics: agent-evals

## Uncertainty

- None recorded.

## 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.