#agent-evaluation

10 Agent Evaluation Techniques Every AI Engineer Should Know (Before Shipping Anything)

Your AI agent worked perfectly yesterday. Today, after a tiny prompt change, it suddenly forgets how to use tools, calls the wrong API, and…

Agent Skill Evals: How Do You Test an AI Agent That Keeps Learning New Skills?

AI evaluation used to have a relatively simple shape.

Harbor Turns Agent Eval Sandboxes Into a Swappable Backend

Running agent evaluations at scale usually means rebuilding your sandbox setup for every provider you touch. Harbor’s pitch: write the eval…

What I Learned Evaluating Cline and Cursor in Production

The lesson that stuck with me came from two failures. A developer points Cline at a feature branch overnight, the autonomous loop runs for…

Careful Adoption of Agentic AI Services: How to Evaluate AI Agents Before They Take Action

Evaluating an AI agent is not the same as reviewing a chatbot response. A chatbot may answer, summarise, or suggest. An AI agent may call a…

The Benchmark Trap: Why AI Agents Look Ready Until You Test Them Like Products

AI agents are having their “show me” moment.

Enterprise AI Evaluation Is Not a Scorecard. It Is a Feedback Flywheel.

Enterprise AI evaluation should not be treated as a scorecard. It should be treated as the operating system for product improvement.

It’s the Rail, Not the Model: A 7-Model Factorial on LLM Financial Hallucination

AI Agent Evaluation Is Broken If You Only Read the Final Answer

How I used Kestra to turn AI agent traces into recurring failure patterns, human-reviewed reports, and an engineering backlog

What Shipping a Browser Agent Taught Me About Evals

Six failure modes broke it in production, none of them on any public benchmark, and here is how I learned to score each one.

Mastering Agent Evaluation. Part 2

Part 2 is devoted to implementing an evaluation system in a project. It provides an overview of the tools and approach, rather than the…

From Prototype to Production: Engineering AI Agents

Building a useful agent is an engineering exercise, not a research project. The bottleneck is not the model. It is the discipline around…

From Prototype to Production: What Changes When You Ship an AI Agent

Your demo worked beautifully. Then you deployed it. Here’s everything that falls apart — and how to fix it before your users find out.