Reduce verbosity in reviewing-changes skill for clean PRs

[SaintPatrck]

🎟️ Tracking

N/A

📔 Objective

This PR reduces excessive verbosity in the reviewing-changes skill when reviewing clean PRs with no issues.

Problem:
PR #6117 received an ~800 token review with elaborate praise sections (Key Strengths, Changes, Code Quality & Architecture, etc.) when no issues were found. The expected output should have been ~30 tokens: "APPROVE - Clean refactoring".

Root Causes:

• No explicit guidance for "clean PR" scenario
• Positive feedback rules understated in skill documentation
• Contradictory verbose example in refactoring checklist
• Insufficient emphasis on anti-patterns for excessive praise

Solution:
Modified 4 skill files to establish "minimal reviews for clean PRs" as the primary principle:

1. SKILL.md: Added "Special Case: Clean PRs" section with explicit 2-3 line approval format; made "minimal reviews" the first core principle
2. review-psychology.md: Elevated brevity guidance to first directive position
3. review-outputs.md: Added Example 7 showing actual PR PM-27756: Create common ExitManager #6117 verbose output as anti-pattern with correct minimal approach
4. refactoring.md: Replaced single verbose example with two distinct examples (with issues vs clean)

Expected Impact:

• 70-90% token reduction for clean PRs (800 → 30-100 tokens)
• Maintains comprehensive reviews for PRs with actual issues
• Reduces noise in PR conversations
• Faster approvals for good work
• Clearer expectations for appropriate review verbosity

Testing:
The skill should now produce minimal ~30 token approvals for clean PRs while maintaining detailed inline comments with collapsible <details> tags for PRs with actual issues.

---

Note: This is a documentation-only change to the .claude/skills/reviewing-changes/ skill. No application code is modified.

---

⏰ Reminders before review

• Contributor guidelines followed
• All formatters and local linters executed and passed
• Written new unit and / or integration tests where applicable
• Protected functional changes with optionality (feature flags)
• Used internationalization (i18n) for all UI strings
• CI builds passed
• Communicated to DevOps any deployment requirements
• Updated any necessary documentation (Confluence, contributing docs) or informed the documentation team

🦮 Reviewer guidelines

• 👍 (:+1:) or similar for great changes
• 📝 (:memo:) or ℹ️ (:information_source:) for notes or general info
• ❓ (:question:) for questions
• 🤔 (:thinking:) or 💭 (:thought_balloon:) for more open inquiry that's not quite a confirmed issue and could potentially benefit from discussion
• 🎨 (:art:) for suggestions / improvements
• ❌ (:x:) or ⚠️ (:warning:) for more significant problems or concerns needing attention
• 🌱 (:seedling:) or ♻️ (:recycle:) for future improvements or indications of technical debt
• ⛏ (:pick:) for minor or nitpick changes

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.89%. Comparing base (0959284) to head (22e9d30).
⚠️ Report is 7 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #6121   +/-   ##
=======================================
  Coverage   84.89%   84.89%           
=======================================
  Files         724      724           
  Lines       52778    52778           
  Branches     7672     7672           
=======================================
  Hits        44807    44807           
  Misses       5279     5279           
  Partials     2692     2692           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[claude]

Claude finished @SaintPatrck's task —— View job

---

Review Complete

Overall Assessment: APPROVE

Excellent skill documentation improvements. The changes effectively address the verbosity issue by establishing "minimal reviews for clean PRs" as a first-class principle with clear examples and anti-patterns.

Analysis:

• SKILL.md: Correctly prioritizes "Minimal reviews for clean PRs" as first principle with explicit guidance
• review-psychology.md: Appropriately elevates brevity to first directive
• review-outputs.md: Strong Example 7 with actual PR PM-27756: Create common ExitManager #6117 as anti-pattern demonstration
• refactoring.md: Good contrast between incomplete migration vs clean refactoring examples

Strengths:

• Consistent messaging across all 4 files
• Concrete examples with token counts (30 vs 800)
• Clear "What NOT to do" sections prevent future mistakes
• Maintains comprehensive review guidance for PRs with actual issues

Minor Observations:

I left 3 inline comments with suggestions for further refinement:

1. Consider consolidating the first 4 principles in SKILL.md to reduce redundancy
2. Add context note explaining why Example 7 was added to review-outputs.md
3. (Acknowledged) Excellent improvement on refactoring examples

These are optional enhancements - the PR is ready to merge as-is.

Expected Impact: Should achieve the stated 70-90% token reduction for clean PRs while maintaining quality for PRs needing detailed feedback.

---

[claude]

ACKNOWLEDGED: Excellent improvement

Details

The addition of Example 2 showing a clean refactoring is exactly what was needed. The explicit "What NOT to do" section with the anti-pattern makes the guidance crystal clear.

The contrast between:

• ✅ "3 lines total"
• ❌ "20 more checkmarks..."

effectively communicates the point without being preachy.

[github-actions]

Checkmarx One – Scan Summary & Details – 447cdcb3-2585-4b96-904f-68e838139742

Great job! No new security vulnerabilities introduced in this pull request