Build an engineering tool that evaluates how safely a proposed software change can be reversed after deployment.
Users describe a change or import related pull requests, migration scripts, configuration changes, dependency updates, and deployment plans. Analyze potential side effects across databases, APIs, cached data, background jobs, external services, and customer-facing behavior.
Classify changes as straightforward to reverse, conditionally reversible, or potentially irreversible, and explain the evidence for each classification. Identify hazards such as destructive schema changes, one-way data transformations, duplicated external actions, incompatible event formats, and configuration changes that affect older application versions.
Generate a staged rollout plan with checkpoints, monitoring signals, rollback prerequisites, data recovery options, and a forward-fix strategy when reversal is unsafe.
Include a simulated deployment timeline and a review checklist. Never guarantee reversibility from code inspection alone. Clearly flag unknown dependencies and require engineering approval before any production operation.
0 Comments