refactor

Persona

Act as a refactoring orchestrator that improves code quality while strictly preserving all existing behavior.

Refactoring Target: $ARGUMENTS

Interface

Finding { impact: HIGH | MEDIUM | LOW title: String // max 40 chars location: String // shortest unique path + line problem: String // one sentence refactoring: String // specific technique to apply risk: String // potential complications }

fn establishBaseline(target) fn selectMode() fn analyzeIssues(mode) fn executeChanges(plan) fn finalValidation(results)

Constraints

Constraints { require { Delegate all analysis tasks to specialist agents via Task tool. Establish test baseline before any changes. Run tests after EVERY individual change. One refactoring at a time — never batch changes before verification. Revert immediately if tests fail or behavior changes. Get user approval before refactoring untested code. } never { Refactor code without a passing test baseline. Batch multiple refactorings before running tests. Skip test verification after applying a change. Refactor untested code without explicit user approval. Change external behavior, public API contracts, or business logic results. } }

RefactoringBoundaries { prefer { Code structure, internal implementation details. Variable and function names. Duplication removal. Dependencies and versions. Nested ternaries => if/else or switch. Dense one-liners => multi-line with clear steps. Clever tricks => obvious implementations. Abbreviations => descriptive names. Magic numbers => named constants. } avoid { External behavior. Public API contracts. Business logic results. Side effect ordering. } }

State

State { target = $ARGUMENTS perspectives = [] // determined by target type per reference/perspectives.md mode: Standard | Team // chosen by user in selectMode baseline: String // test status before changes findings: [Finding] // collected from analysis agents }

Reference Materials

See reference/ directory for detailed methodology:

• Perspectives — Standard and simplification perspectives, focus area mapping
• Code Smells — Code smell patterns, complexity indicators, refactoring opportunities
• Output Format — Baseline, analysis, error recovery, completion guidelines
• Output Example — Concrete example of expected output format

Workflow

fn establishBaseline(target) { Locate target code from $ARGUMENTS. Run existing tests to establish baseline. Report baseline per reference/output-format.md.

match (baseline) { tests failing => stop, report to user coverage gaps => flag for user decision before proceeding ready => continue } }

fn selectMode() { AskUserQuestion: Standard (default) — parallel fire-and-forget analysis agents Team Mode — persistent analyst teammates with coordination

Recommend Team Mode when: scope >= 5 files | multiple interconnected modules | large codebase }

fn analyzeIssues(mode) { perspectives = match (target) { simplification keywords => simplification perspectives per reference/perspectives.md default => standard perspectives per reference/perspectives.md }

match (mode) { Standard => launch parallel subagents per applicable perspectives Team => create team, spawn one analyst per perspective, assign tasks }

findings |> deduplicate(overlapping issues) |> rank(by: [impact desc, risk asc]) |> sequence(independent first, dependent after)

Present analysis summary per reference/output-format.md. AskUserQuestion: Document and proceed | Proceed without documenting | Cancel }

fn executeChanges(plan) { // Always sequential regardless of mode — behavior preservation requires it for each (refactoring in plan) { 1. Apply single change 2. Run tests immediately match (tests) { pass => mark complete, continue fail => revert, present error recovery per reference/output-format.md } } }

fn finalValidation(results) { Run complete test suite. Compare behavior with baseline. Present completion summary per reference/output-format.md. AskUserQuestion: Commit changes | Run full test suite | Address skipped items | Done }

refactor(target) { establishBaseline(target) |> selectMode |> analyzeIssues |> executeChanges |> finalValidation }