| id | hatch3r-logical-refactor |
| name | hatch3r-logical-refactor |
| type | skill |
| description | Workflow for changing behavior or logic flow without adding new features or overhauling UI. Use when modifying business logic, data flows, behavioral rules, or working on logical refactor issues. |
| tags | ["implementation"] |
| quality_charter | agents/shared/quality-charter.md |
| efficiency_patterns | agents/shared/efficiency-patterns.md |
| cache_friendly | true |
Note: Commands below use npm as an example. Substitute with your project's package manager (yarn, pnpm, bun) or build tool when your project uses a different package manager.
Logical Refactor Workflow
Quick Start
Task Progress:
- [ ] Step 0: Detect ambiguity (P8 B1)
- [ ] Step 1: Read the issue, specs, and existing tests
- [ ] Step 2: Produce a change plan
- [ ] Step 3: Implement the behavior change
- [ ] Step 4: Update tests and verify invariants
- [ ] Step 5: Open PR
Step 0 — Detect Ambiguity (P8 B1)
Before any work, scan the invocation for unresolved questions in scope, intent, acceptance criteria, target environment, or irreversibility. If any are found, ask the user via the platform-native question tool per agents/shared/user-question-protocol.md. Do not proceed under silent assumption. Default path, not an exception. Triggers for THIS skill: invariants to preserve vs change, before/after behavior specification, downstream consumer impact, spec update authority (this PR vs follow-up), and characterization-test requirement when current behavior is undertested.
Fan-out Discipline (P8 B2)
Fan-out scales with task size; token cost never justifies serializing independent work (rules/hatch3r-fan-out-discipline.md P8 B2; agents/shared/efficiency-patterns.md). Emit sub_agents_spawned: { count, rationale, task_structure } in your output.
Step 1: Read Inputs
- Parse the issue body: motivation, before/after behavior, invariants preserved, invariants changed, acceptance criteria, affected files, risk analysis, testing plan.
- Read relevant project documentation: behavior engine, event model, data model, privacy, quality.
- Review existing tests to understand what behavior is currently asserted.
- For external library docs and current best practices, follow the project's tooling hierarchy.
Step 2: Change Plan
Before modifying code, output:
- Current behavior: how it works now
- New behavior: how it will work after
- Invariants preserved: list from issue + specs
- Invariants changed: list, with justification
- Cascading effects: which other flows/components are affected
- Files to modify: list with what changes
- Risks: what could go wrong
Step 3: Implement
- Implement behavior change as described in the before/after section.
- Verify every listed invariant is maintained (or explicitly updated with justification).
- Update any spec docs affected by the behavior change.
Step 4: Test
- Update existing tests that assert old behavior to assert new behavior.
- Add new tests for the new behavior, especially edge cases.
- Regression tests for preserved invariants.
- Integration tests if change affects cross-module interactions.
${HATCH3R:VERIFY_GATE_ALL}
Resolved to the project's language-aware gate at sync time (fallback when detection is unknown: npm run lint && npm run typecheck && npm run test).
Step 5: Open PR
Use the project's PR template. Include:
- Before/after behavior with concrete examples
- Invariants preserved and changed
- Test evidence for both new behavior and preserved behavior
- Spec docs updated (if any)
Error Handling
- Behavioral invariant violated during refactor: If a test that should pass now fails, check whether the invariant was incorrectly preserved or whether the refactor inadvertently changed behavior. Fix the refactor, not the test, unless the test was wrong.
- Refactor scope grows beyond the original task: If additional modules need changes to complete the refactor, stop, document the expanded scope, and get confirmation before continuing.
- Cannot verify behavioral equivalence due to missing tests: Write characterization tests for the current behavior before applying the refactor. This provides a safety net for the transformation.
Definition of Done