implement

To convert a task description into working code artifacts, provide a plain text task or issue URL with optional context references like research.md or plan.md. The system normalizes the input, extracts success criteria, and generates concrete code through a structured quality loop.

Yes, you can generate code from a GitHub, Jira, or Linear issue URL. The implementation pipeline accepts issue URLs as task sources, normalizing them to extract what needs to change, why, and the success criteria before producing the code artifact.

The quality loop enforces standards during code generation by applying mechanical checks and an intent review. This ensures the generated implementation artifact meets the extracted success criteria before producing the final implementation outputs.

No, the code generation pipeline does not automatically create pull requests. It focuses on producing implementation artifacts and quality documentation while avoiding live QA or automated PR creation, leaving integration to downstream stages or the orchestrator.

You can use research.md, debug.md, and plan.md as context artifacts for software implementation tasks. These references provide necessary background for the implementation pipeline to understand the task and generate the working code artifact.