An agent unable to build and test your project cannot verify its own work. Everything that follows — quality checks, reviews, and the confidence you place in a pull request — depends on these commands being correct.
Coroid identifies these when a project is created. Check what it has found.
What Coroid requires
| Command | Used for |
|---|---|
| Install | Restoring dependencies in a clean workspace |
| Build | Confirming the project compiles |
| Test | Running your test suite and viewing the results |
The workspace is clean and disposable per task, so every command must function from a fresh clone with no cached files and nothing pre-installed.
Where detection often goes awry
Monorepos. Commands reside in a subdirectory or require a workspace filter. The root-level command might exist but produce incorrect results.
Setup steps. Test suites needing code generation, a migration, or a fixture build before they can run. Detection spots the test command but overlooks its prerequisites.
Service dependencies. Tests requiring a database, cache, or queue to be running.
If your suite assumes localhost:5432 is active, it will fail in a clean workspace.
Environment variables. Test suites reading configuration from an .env that is not stored in the repository.
Signs of a faulty configuration
The symptom is tasks failing during verification for reasons unrelated to the change — the same error occurring on every task at the same stage, regardless of what was requested.
If the first error you see occurs before any code is written, it is almost always the build command. If it appears after the change is made, it is usually the test command or something the tests depend on.
Containers
Where a project can be containerised, Coroid can use that as the execution environment — resolving service dependencies and environment drift in one go, as the container already includes them.
If you have a functional Dockerfile or Compose file that runs your tests, that is typically the most reliable configuration available to you.
Language support
Agents operate in any language, but the extent to which Coroid can verify things independently varies — particularly semantic navigation. See language support and the agent environment to learn what is installed in the workspace and which languages have a language server.
Getting these commands right is more important than the language you use. A correct build and test pair in a poorly supported language outperforms a faulty one in a well-supported language.