Fix RuboCop Violations Skill

Assists with systematically reducing RuboCop offenses while preserving behavior and keeping changes focused.

IMPORTANT: When using this skill, announce to the user: "Using fix-rubocop skill to fix RuboCop violations systematically."

Goals

• Keep changes minimal and behavior-preserving: Prefer small, focused changes
• Use safe autocorrect first: Apply unsafe corrections only when reviewed and covered by tests
• Centralize rule changes in .rubocop.yml: Never hand-edit .rubocop_todo.yml
• No inline disables: This project forbids # rubocop:disable comments

Quick Reference Commands

# Full lint check
rake rubocop

# Lint specific file or directory
bundle exec rubocop path/to/file.rb

# Safe autocorrect (recommended first step)
bundle exec rubocop -a path/to/file.rb

# Unsafe autocorrect for specific cop (use with caution)
bundle exec rubocop -A --only Cop/Name path/to/file.rb

# Run tests to verify no behavior changes
rake spec

# Regenerate .rubocop_todo.yml after fixes
bundle exec docquet regenerate-todo

Complete Workflow

Step 1: Reproduce and Identify Offenses

# Run full lint to see all offenses
rake rubocop

# Note the failing cops and affected files
# Example output:
# lib/foxtail/formatter.rb:15:3: C: Style/StringLiterals: ...

Identify the scope:

• Which cop(s) are failing?
• Which file(s) are affected?
• How many offenses?

Step 2: Safe Autocorrect First

Start with the safest approach - automatic fixes that preserve behavior:

# Target specific files to reduce blast radius
bundle exec rubocop -a lib/foxtail/specific_file.rb

# Or target a directory
bundle exec rubocop -a lib/foxtail/

# Verify no new failures
rake rubocop

# CRITICAL: Verify tests still pass
rake spec

If tests fail after autocorrect:

• Revert changes: git checkout lib/foxtail/specific_file.rb
• Proceed to Step 3 for manual fixes

Step 3: Target Specific Cops (If Still Failing)

For remaining offenses, fix specific cops one at a time:

# For stylistic issues
bundle exec rubocop -A --only Style/StringLiterals lib/foxtail/

# For performance issues (manual review recommended)
bundle exec rubocop --only Performance/MapCompact lib/foxtail/
# Review and fix manually based on suggestions

# Verify after each cop
rake rubocop
rake spec

Important: For logic-sensitive cops (Performance, Lint), prefer manual fixes with test verification.

Step 3a: Fixing Cops Disabled in .rubocop_todo.yml

If a cop is disabled in .rubocop_todo.yml, you must temporarily enable it:

# 1. Check if cop is in TODO file
grep -A 5 "Cop/Name" .rubocop_todo.yml

# 2. Temporarily comment out that cop's block in .rubocop_todo.yml
# (RuboCop respects TODO disables even with --only)

# 3. Run targeted fix
bundle exec rubocop --only Cop/Name path/to/file.rb
# or with autocorrect:
bundle exec rubocop -A --only Cop/Name path/to/file.rb

# 4. Validate changes
rake rubocop
rake spec

# 5. Regenerate TODO file (removes fixed cops automatically)
bundle exec docquet regenerate-todo

# 6. Commit code changes AND regenerated .rubocop_todo.yml together

CRITICAL: Always use docquet regenerate-todo to update the TODO file. Never edit it manually.

Step 4: Update Configuration (Rare)

If a rule conflicts with project style:

# Edit .rubocop.yml to adjust the rule
# Add clear comment explaining the rationale

Example in .rubocop.yml:

# We prefer double quotes for consistency with string interpolation
Style/StringLiterals:
  EnforcedStyle: double_quotes

After configuration changes:

# Regenerate TODO to reflect new config
bundle exec docquet regenerate-todo

# Verify
rake rubocop

Step 5: No Inline Disables

FORBIDDEN:

# ❌ NEVER do this
def some_method
  # rubocop:disable Style/SomeCop
  code_here
  # rubocop:enable Style/SomeCop
end

If a finding cannot be safely fixed:

1. Ask the user for guidance
2. Options:
   • Refine the rule in .rubocop.yml
   • Adjust the code approach
   • Document why this specific pattern is needed

Step 6: Validate and Commit

# Ensure everything is clean
rake

# Should show:
# - All tests passing
# - No RuboCop offenses (or reduced offenses)

Commit message format:

git commit -m "$(cat <<'EOF'
:police_officer: Fix Style/StringLiterals violations in lib/foxtail

Convert single quotes to double quotes for consistency.

:robot: Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
EOF
)"

Commit message patterns:

• :police_officer: - All RuboCop fixes (style, lint, performance)
• :wrench: - Update RuboCop configuration

Common Scenarios

Scenario 1: Reduce .rubocop_todo.yml Violations

# 1. Check current state
cat .rubocop_todo.yml
# Note: List of disabled cops and offense counts

# 2. Pick one cop to fix (start with lowest count)
# Example: Style/CommentedKeyword (1 offense)

# 3. Find affected files
bundle exec rubocop --only Style/CommentedKeyword

# 4. Apply fix
bundle exec rubocop -A --only Style/CommentedKeyword lib/foxtail/

# 5. Verify
rake spec
rake rubocop

# 6. Regenerate TODO
bundle exec docquet regenerate-todo
# The fixed cop should be removed automatically

# 7. Commit
git add .
git commit -m ":police_officer: Fix Style/CommentedKeyword violations"

Scenario 2: Fix All Offenses in One File

# 1. Target specific file
bundle exec rubocop lib/foxtail/date_time_format.rb

# 2. Safe autocorrect first
bundle exec rubocop -a lib/foxtail/date_time_format.rb

# 3. Check remaining offenses
bundle exec rubocop lib/foxtail/date_time_format.rb

# 4. Fix remaining manually or with targeted cops
bundle exec rubocop -A --only Naming/VariableNumber lib/foxtail/date_time_format.rb

# 5. Verify
rake spec

# 6. If cop was in TODO, regenerate
bundle exec docquet regenerate-todo

# 7. Commit
git commit -m ":police_officer: Fix all RuboCop violations in date_time_format.rb"

Scenario 3: Fix Specific Cop Across Entire Codebase

# 1. Check scope
bundle exec rubocop --only Style/StringLiterals
# Shows all files and offense counts

# 2. Start with smallest files or directories
bundle exec rubocop -A --only Style/StringLiterals lib/foxtail/helpers/

# 3. Verify incrementally
rake spec
git add lib/foxtail/helpers/
git commit -m ":police_officer: Fix Style/StringLiterals in helpers"

# 4. Continue with other directories
bundle exec rubocop -A --only Style/StringLiterals lib/foxtail/models/
rake spec
git commit -m ":police_officer: Fix Style/StringLiterals in models"

# 5. Final regenerate
bundle exec docquet regenerate-todo
git commit -m ":wrench: Regenerate RuboCop TODO after fixes"

Examples by Cop Type

Style Cops (Usually Safe)

# String literals
bundle exec rubocop -A --only Style/StringLiterals lib/

# Hash syntax
bundle exec rubocop -A --only Style/HashSyntax lib/

# Frozen string literals
bundle exec rubocop -A --only Style/FrozenStringLiteralComment lib/

Naming Cops (May Need Review)

# Variable naming
bundle exec rubocop -A --only Naming/VariableNumber lib/foxtail/file.rb

# Method naming (often requires manual decision)
bundle exec rubocop --only Naming/PredicateMethod lib/foxtail/file.rb
# Review output and fix manually:
# - Either rename method to end with ?
# - Or document why it shouldn't be a predicate

Lint Cops (Require Careful Review)

# Unused variables (manual fix recommended)
bundle exec rubocop --only Lint/UnusedMethodArgument lib/
# Review each and either use or prefix with _

# Verify with tests
rake spec

Performance Cops (Test Thoroughly)

# Performance improvements (review before applying)
bundle exec rubocop --only Performance/MapCompact lib/
# Apply manually to ensure behavior preservation

# ALWAYS run full test suite
rake spec

Common Pitfalls to Avoid

1. Don't run unsafe autocorrect on entire codebase: Use -A on small scopes only
2. Don't skip test verification: Always run rake spec after fixes
3. Don't manually edit .rubocop_todo.yml: Always use docquet regenerate-todo
4. Don't use inline disables: This project forbids # rubocop:disable comments
5. Don't commit without regenerating TODO: If fixing TODO cops, always regenerate
6. Don't mix unrelated fixes: Keep commits focused on one cop or one area
7. Don't ignore test failures: Revert and fix manually if tests fail after autocorrect

Troubleshooting

Issue: Cop not showing offenses despite being in .rubocop_todo.yml

Problem: Cop is disabled in TODO file

Solution:

# Temporarily comment out the cop in .rubocop_todo.yml
# Then run:
bundle exec rubocop --only Cop/Name

# After fixing, regenerate TODO
bundle exec docquet regenerate-todo

Issue: Tests fail after autocorrect

Problem: Unsafe autocorrect changed behavior

Solution:

# Revert changes
git checkout path/to/file.rb

# Fix manually or use safer approach
bundle exec rubocop --only Cop/Name path/to/file.rb
# Review suggestions and fix by hand

# Verify each change
rake spec

Issue: docquet regenerate-todo not working

Problem: Command not found or gem not installed

Solution:

# Ensure bundler is up to date
bundle install

# Check if docquet is installed
bundle exec docquet --version

# If not, check Gemfile and install dependencies

Issue: Too many offenses to fix at once

Problem: Overwhelming number of violations

Solution:

1. Pick ONE cop type to fix
2. Start with lowest offense count
3. Fix incrementally, one file/directory at a time
4. Commit after each successful fix
5. Use PR per cop type or per area

PR Guidance

Keep changes small and single-purpose:

• Good: "Fix Style/StringLiterals in lib/foxtail/formatters"
• Bad: "Fix multiple RuboCop issues across codebase"

Include context in PR body:

## Summary
Fix Style/StringLiterals violations in formatter classes

## Changes
- Convert single quotes to double quotes in 5 files
- Applied safe autocorrect (-a flag)

## Verification
- All tests pass (rake spec)
- RuboCop clean for these files
- Regenerated .rubocop_todo.yml

:robot: Generated with [Claude Code](https://claude.com/claude-code)

Link rule changes to rationale: If updating .rubocop.yml, explain why in PR description.

Best Practices

1. Start with safe autocorrect: Use -a before -A
2. Small, focused PRs: One cop type or one directory at a time
3. Test after every change: Run rake spec frequently
4. Regenerate TODO incrementally: After fixing each cop type
5. Commit incrementally: Don't wait to fix everything before committing
6. Document decisions: If excluding a cop, explain why in .rubocop.yml
7. Review Performance/Lint changes: Don't blindly autocorrect these
8. Keep blast radius small: Target specific files/directories when possible