What problem does it solve? Enforcing consistent error management across a large Go repository is tedious and error-prone: bare panics, raw error returns, and ad-hoc multi-value returns slip through code review. This Skill autonomously audits, refactors, and validates repository-wide error handling against the rules defined in 02-spec/03-error-manage/. ## Core Features & Use Cases - Invariant Enforcement: Detects and eliminates bare panics, os.Exit calls, and raw standard library error returns, replacing them with structured *appfault.AppError wrappers carrying Op, Code, Type, Severity, and Cause context. - Result Envelope Refactoring: Converts multi-value returns like (T, error) into typed appfault.Result[T], ResultSlice[T], or ResultMap[K, V] containers with pointer-receiver predicate methods such as IsFailure, IsEmpty, and IsCountOtherThan. - Targeted Verification: Validates only modified files via linter-scripts/check-error-management.py instead of running the full CI/CD pipeline, and records changes atomically in .ai-memory/temp/recent-file-changes.json. - Use Case: After adding a new Go package, run this Skill to sweep the package for banned error patterns, refactor returns into Result envelopes, define reusable types in types.go, and commit everything in a single atomic push. ## Quick Start Audit the Go files in my new package for banned error patterns, refactor them to use AppError wrappers and Result envelopes, verify with the targeted linter, and commit the changes.
