feat(pipeline): split PreFlight into granular setup steps (#302)

[ArthurCRodrigues]

Context

The current PreFlightStep was a "god step" that handled file checks, asset injection, and setup commands all at once. This made it difficult to track granular failures and introduced coupling between different setup phases.

Solution

• Split PreFlightStep into three single-purpose steps: FileCheckStep, AssetInjectionStep, and SetupCommandsStep.
• Updated Step abstract class to include a category property for better organization.
• Updated Template abstract class to allow templates to contribute setup requirements (required files and setup commands).
• Implemented merging logic in build_pipeline to combine assignment-level and template-level setup configurations.
• Updated PipelineExecutionSerializer and feedback_generator to support the new granular steps and provide detailed success/failure reporting.
• Updated StepRegistry to handle the new steps and their conditional inclusion logic.
• Updated SandboxManager to handle Docker client initialization lazily, allowing the API to run in remote mode without requiring a local Docker socket.

Further clarifications

• PreFlightStep class is still kept for compatibility but is no longer used in the default pipeline orchestration.
• Pipeline execution summary now shows a more detailed breakdown of the setup phase.
• API consumers should update their logic to handle the new step names if they were specifically looking for PreFlightStep.

Related issues

Closes #302

Checklist

• I linked the related issue(s) and explained the motivation.
• I kept this PR focused and scoped to a single concern.
• I added or updated tests for changed behavior (or explained why not needed).
• I ran the relevant tests locally.
• I updated documentation when needed (README/docs/API examples).
• This PR introduces API contract changes (request/response/endpoint/DTO).
• If API changed, I documented compatibility or migration notes.
• This PR includes breaking changes.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.