# Miroir - [Introduction](https://miroir.mintlify.app/index.md): Turn your product into a Harbor environment. - [Quick start](https://miroir.mintlify.app/quick-start.md): Build a small CRM world, test its graders through Harbor, and read what a trial recorded. - [Core concepts](https://miroir.mintlify.app/core-concepts.md): What happens during a trial, and the few terms the rest of the docs rely on. - [Wrap a product](https://miroir.mintlify.app/environment.md): Turn your product's container image into a world image that Harbor can run and Miroir can grade. - [Write tasks](https://miroir.mintlify.app/tasks.md): The task directory, its instruction, reference solution and probes, and how to get from reference runs to a grading spec. - [Validate graders](https://miroir.mintlify.app/validate.md): Hold every grader to its reference solution, its null run and its probes before any agent runs on it. - [Publish](https://miroir.mintlify.app/publish.md): Pin the images, pass the release gate, and publish the dataset with Harbor. - [Train](https://miroir.mintlify.app/train.md): The reward contract, how to pick the tasks worth training on, and the trainers that run Miroir tasks. - [Manifest](https://miroir.mintlify.app/manifest.md): Every key of /miroir/miroir.yaml, and how dates, sinks and quiet work. - [World lifecycle](https://miroir.mintlify.app/lifecycle.md): What Miroir does at build, at each boot and at collect, what keeps the agent out, and what each failure message means. - [Grading spec](https://miroir.mintlify.app/spec.md): The MIROIR_SPEC format, how checks, gates and allow entries score, every outcome, and the rules for writing a spec. - [Evidence](https://miroir.mintlify.app/evidence.md): The tables a check can read, how their values look, and recipes for common checks. ## Optional - [Harbor docs](https://docs.harborframework.com) This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.