{"slug": "how-to-set-up-a-project-to-let-agents-validate-work", "title": "How to set up a project to let agents validate work", "summary": "Software engineer and DeepFocusTimer creator shared a prompt to help developers set up projects so AI agents can validate their own work by running tests and using the app, addressing the challenge of closing the feedback loop in agentic software engineering. The prompt instructs agents to document how to run tests and interact with the application, improving agent effectiveness.", "body_md": "The biggest benefits of agentic software engineering are hidden behind one key challenge.\n\nYour agents **have to** be able to validate their own work.\n\nAgents aren’t super useful without being able to close the feedback loop on their work. They’ll write code, but then you have to poke around your app and paste any feedback. It’s a waste of time.\n\nBut that’s how software is made. You write some code. You run some tests. You iterate.\n\nYou wouldn’t write and ship code without trying to validate that it works, and you shouldn’t do it with agents either.\n\nBut your agents should actually be able to run tests, and ideally, use the app.\n\n## Why many projects fall short\n\nClosing the feedback loop is often done in two ways:\n\nRunning tests (like unit tests)\n\nPerforming actions against your application\n\nFor an API, performing actions might be making API calls. Maybe it’s automated, maybe it’s Postman/cURL.\n\nBut for web apps or mobile apps, it gets more complicated. For [DeepFocusTimer](http://deepfocustimer.com), I run a development server, visit the app in my browser, and manually validate changes. Most apps, it’s a bit of both.\n\nSo most of the time, agents need to be able to:\n\nRun tests\n\nManually use the app\n\nThe first of these is easy. Agents can often figure out how to run tests, but apps that have explicit instructions on how to run tests in their `AGENTS.md`\n\nor `CLAUDE.md`\n\nget better results.\n\n## A prompt to get your project ready for agent-validation\n\nYou can fix all of the above problems manually, but I have a prompt that I’ve been using to fix all my projects. Here it is:", "url": "https://wpnews.pro/news/how-to-set-up-a-project-to-let-agents-validate-work", "canonical_source": "https://www.augmentedswe.com/p/how-to-set-up-a-project-to-let-agents", "published_at": "2026-08-20 12:24:06+00:00", "updated_at": "2026-08-20 12:42:57.327546+00:00", "lang": "en", "topics": ["ai-agents", "developer-tools"], "entities": ["DeepFocusTimer"], "alternates": {"html": "https://wpnews.pro/news/how-to-set-up-a-project-to-let-agents-validate-work", "markdown": "https://wpnews.pro/news/how-to-set-up-a-project-to-let-agents-validate-work.md", "text": "https://wpnews.pro/news/how-to-set-up-a-project-to-let-agents-validate-work.txt", "jsonld": "https://wpnews.pro/news/how-to-set-up-a-project-to-let-agents-validate-work.jsonld"}}