QA Refactoring Safety

Use this skill to refactor safely: preserve behavior, reduce risk, and keep CI green while improving maintainability and delivery speed.

Defaults: baseline first, smallest safe step next, and proof via tests/contracts/observability instead of intuition.

Quick Start (10 Minutes)

• Confirm baseline: main green; reproduce the behavior you must preserve.
• Choose a boundary: API surface, module boundary, DB boundary, or request handler.
• Add a safety net: characterization/contract/integration tests at that boundary.
• Refactor in micro-steps: one behavior-preserving change per commit/PR chunk.
• Prove: run the smallest relevant suite locally, then full CI; keep failures deterministic.

Core QA (Default)

Safe Refactor Loop (Behavior First)

• Establish baseline: get main green; reproduce the behavior you must preserve.
• Define invariants: inputs/outputs, error modes, permissions, data shape, performance budgets.
• Add a safety net: write characterization/contract/integration tests around the boundary you will touch.
• Create seams: introduce injection points/adapters to isolate side effects and external dependencies.
• Refactor in micro-steps: one behavior-preserving change at a time; keep diffs reviewable.
• Prove: run the smallest relevant suite locally, then full CI; keep failures debuggable and deterministic.
• Ship safely: use canary/dark launch/feature flags when refactors touch production-critical paths.

Risk Levels (Choose Safety Net)

Test Strategy for Refactors

• Prefer contract and integration tests around boundaries to preserve behavior.
• Use snapshots/golden masters only when outputs are stable and reviewed (avoid "approve everything" loops).
• For invariants, consider property-based tests or table-driven cases (inputs, edge cases, error modes).
• Avoid making E2E/UI tests the primary safety net for refactors; keep most safety below the UI.
• For flaky areas: fix determinism first (seeds, time, ordering, network) before trusting results.

CI Economics and Debugging Ergonomics

• Keep refactor PRs small and reviewable; avoid refactor + feature in one PR.
• Require failure artifacts for tests guarding refactors (logs, trace IDs, deterministic seeds, repro steps).
• Reduce diff noise: isolate formatting-only changes (or apply formatting repo-wide once with buy-in).
• Keep git bisect viable: avoid mixed "mechanical + semantic" changes unless necessary.

Do / Avoid

Do:

• Add missing tests before refactoring high-risk areas.
• Add guardrails (linters, type checks, contract checks, static analysis/security checks) so refactors don't silently break interfaces.
• Prefer "branch by abstraction" / adapters when you need to swap implementations safely.

Avoid:

• Combining large structural refactors with behavior changes.
• Using flaky E2E as the primary safety net for refactors.

Quick Reference

Decision Tree: Refactoring Strategy

Code issue: [Refactoring Scenario]
    ├─ Code Smells Detected?
    │   ├─ Duplicated code? → Extract method/function
    │   ├─ Long method (>50 lines)? → Extract smaller methods
    │   ├─ Large class (>300 lines)? → Split into focused classes
    │   ├─ Long parameter list? → Parameter object
    │   └─ Feature envy? → Move method closer to data
    │
    ├─ Legacy Code (No Tests)?
    │   ├─ High risk? → Write characterization tests first
    │   ├─ Large rewrite needed? → Strangler Fig (incremental migration)
    │   ├─ Unknown behavior? → Characterization tests + small refactors
    │   └─ Production system? → Canary deployments + monitoring
    │
    ├─ Quality Standards?
    │   ├─ New project? → Setup linter + formatter + quality gates
    │   ├─ Existing project? → Add pre-commit hooks + CI checks
    │   ├─ Complexity issues? → Set cyclomatic complexity limits (<10)
    │   └─ Technical debt? → Track in register, 20% sprint capacity

Related Skills

• Debugging production issues: qa-debugging
• Code review process and checklists: software-code-review
• New architecture design from scratch: software-architecture-design
• Test strategy and coverage planning: qa-testing-strategy

Operational Deep Dives

Shared Foundation

• ../software-clean-code-standard/references/clean-code-standard.md - Canonical clean code rules (CC-*) for citation
• Legacy playbook: ../software-clean-code-standard/references/code-quality-operational-playbook.md - RULE-01–RULE-13, decision trees, and operational procedures
• ../software-clean-code-standard/references/refactoring-operational-checklist.md - Refactoring smell-to-action mapping, safe refactoring guardrails
• ../software-clean-code-standard/references/working-effectively-with-legacy-code-operational-checklist.md - Seams, characterization tests, incremental migration patterns

Skill-Specific

See references/operational-patterns.md for detailed refactoring catalogs, automated quality gates, technical debt playbooks, and legacy modernization steps.

Templates

Use copy-paste templates in assets/ for checklists and quality-gate configs:

• Refactoring: assets/process/refactoring-checklist.md, assets/process/code-review-quality.md
• Technical debt: assets/tracking/tech-debt-register.md
• Quality gates: assets/quality-gates/javascript/eslint-config.js, assets/quality-gates/platform-agnostic/sonarqube-setup.md

Resources

Use deep-dive guides in references/ (load only what you need):

• Operational Patterns: references/operational-patterns.md - Core refactoring catalogs, quality gates, and legacy modernization
• Refactoring Catalog: references/refactoring-catalog.md
• Code Smells Guide: references/code-smells-guide.md
• Technical Debt Management: references/tech-debt-management.md
• Legacy Code Modernization: references/legacy-code-strategies.md

Optional: AI / Automation

Do:

• Use AI to propose mechanical refactors (rename/extract/move) only when you can prove behavior preservation via tests and contracts.
• Use AI to summarize diffs and risk hotspots; verify by running targeted characterization tests.
• Prefer tool-assisted refactors (IDE/compiler-aware, codemods) over freeform text edits when available.

Avoid:

• Accepting refactors that change behavior without an explicit requirement and regression tests.
• Letting AI "fix tests" by weakening assertions to make CI green.

See data/sources.json for curated external references.