Cross-Repo Integration Mapping Prompt for Claude

The prompt forces Claude to treat both codebases as first-class inputs simultaneously, rather than analyzing them in isolation. By loading both contexts together, the model can surface mismatches in data models, naming conventions, and interface contracts that would be invisible if each repo were reviewed separately. The explicit list of deliverables — API endpoints, shared data structures, migration strategy, and breakage analysis — prevents vague or partial output. Each item corresponds to a distinct phase of real integration work, so the response maps directly onto actionable engineering tasks rather than high-level observations. Asking "what breaks if we connect them" is particularly effective because it primes the model to reason adversarially rather than optimistically. This shifts the output from aspirational architecture toward a risk-aware plan, surfacing versioning conflicts, incompatible authentication schemes, or circular dependencies before any code is written.