refactor

Activate When

• /godmode:refactor, "refactor this", "clean up"
• "extract", "inline", "move", "rename", "reorganize"
• Review skill identifies maintainability score < 6/10

Workflow

1. Assess Refactoring Scope

find . -name "*.ts" -o -name "*.js" -o -name "*.py" \
  | xargs wc -l | sort -rn | head -20
grep -rn "<pattern>" --include="*.ts" --include="*.js"

Target: <file or module>
Lines: <N> | Complexity: <cyclomatic> | Coverage: <N%>
Dependents: <N files> | Risk: LOW|MEDIUM|HIGH|CRITICAL

Risk levels:

• LOW: <10 dependents, >80% coverage, isolated
• MEDIUM: 10-30 dependents, 50-80% coverage
• HIGH: 30+ dependents, <50% coverage
• CRITICAL: core module, <30% coverage, many dependents

IF coverage < 60%: write characterization tests first. IF dependents > 30: use strangler pattern migration.

2. Select Refactoring Pattern

Extract (Function, Class, Interface, Module, Variable, Parameter), Inline (Function, Variable), Move (Function, Field), Rename (Symbol, File), Simplify (Conditional, Loop, Guard Clause), Replace (Inheritance with Composition, Conditional with Polymorphism).

3. Impact Analysis

grep -rl "<symbol>" --include="*.ts" --include="*.js"

Map directly affected (modified), indirectly affected (imports changed), and not affected files.

4. Verify Safety Net

npm test 2>&1 | tail -10
npx jest --testPathPattern="<target>" 2>&1

IF coverage < 60%: STOP, write characterization tests, commit them separately, then proceed.

5. Execute Refactoring

One transformation per commit. For each step:

1. Apply transformation
2. Run full test suite
3. Pass -> commit: "refactor: <pattern> -- <desc>"
4. Fail -> revert and investigate

IF tests fail after transformation: revert first, then diagnose. Never debug forward.

6. Migration Strategy (Large Refactors)

Strangler pattern: create new alongside old, migrate dependents one at a time, each migration = separate commit, remove old code only after zero references.

IF >5 dependents: phase the migration over multiple PRs.

7. Post-Refactoring Verification

npm test

Compare: test count same or higher, coverage same or higher, no behavior change. Report pattern used, commits, files modified/created/deleted.

Hard Rules

1. NEVER refactor without green test suite first.
2. NEVER combine refactoring with behavior changes.
3. ONE pattern per commit.
4. NEVER skip impact analysis.
5. NEVER delete code without verifying zero references.
6. ALWAYS run tests after every transformation.
7. NEVER force-push refactoring branches.
8. IF tests fail: REVERT first, THEN diagnose.
9. Coverage MUST NOT decrease after refactoring.
10. EVERY renamed symbol: update comments, docs, errors.

Complexity Thresholds

Cyclomatic: <= 10/function (eslint, radon, gocyclo)
Cognitive:  <= 15/function (SonarQube, eslint sonarjs)
LOC:        <= 50/function (wc -l)

TSV Logging

Append .godmode/refactor.tsv:

timestamp	target	pattern	tests_before	tests_after	coverage_before	coverage_after	status

Keep/Discard

KEEP if: tests pass AND complexity reduced
  AND coverage maintained or increased.
DISCARD if: tests fail OR complexity unchanged
  OR coverage dropped.

Stop Conditions

STOP when FIRST of:
  - All targets below thresholds
    (cyclomatic <= 10, cognitive <= 15)
  - Test count and coverage same or higher
  - Zero dead code detected

Autonomous Operation

On failure: git reset --hard HEAD~1. Never pause.

Error Recovery

Failure	Action
Tests fail after transform	Revert, plan smaller step
Coverage decreases	Write tests before continuing
Circular dependency	Extract shared into new module