Toolchain and dependency recovery
Node runtime, lockfile, package manager, framework version, native modules, generated code, and environment variables must describe one reproducible build.
A stalled repository is an evidence problem
Get the app shipping again.
React and Next.js rescue work starts where generic “web repair” stops: package and runtime compatibility, App Router or Pages Router behavior, server and client boundaries, data fetching, environment assumptions, build output, deployment configuration, and a handoff another engineer can actually use.
Submit the failure briefRecovery tracks
Fix the blocking system, not the loudest error.
Server and client boundary
Hydration mismatches, browser-only dependencies, cache behavior, dynamic routes, metadata, and server actions require different evidence than a component styling defect.
Deployment parity
A local success is not production proof. We compare build commands, runtime, environment, routing, headers, storage, third-party services, and observability.
Maintainable handoff
The result includes what failed, why the intervention is bounded, how it was tested, what remains risky, and how to reverse it.
Especially useful for
Inherited and AI-assisted codebases
Fast-generated applications often contain a mix of working UI, duplicated patterns, hidden environment assumptions, incomplete authorization, and dependencies chosen without a release model. The answer is not to shame the source. It is to establish authority, prioritize user-critical paths, and reduce the codebase to testable contracts.