Code Integration

When multiple people contribute to a project built through synthesis coding, integration is fundamentally different from a standard open source merge workflow. This skill defines how it works.

The Core Problem

In synthesis coding, the lead developer (the "lead synthesist") builds and evolves the system through continuous, context-rich collaboration with AI. The result is a codebase with:

• Deep architectural consistency — decisions compound across sessions
• Implicit conventions — not all standards are documented yet because one person held them all in their head
• Rapid evolution — the codebase may change substantially between the time an external contributor branches off and the time they submit a PR

When an external contributor forks or branches, they get a snapshot. They build against that snapshot. Meanwhile, the lead may have evolved the architecture, introduced new patterns, improved output quality, or refactored entire subsystems. The contributor's code reflects the old state.

A blind merge risks: