review-comments（审阅意见）

Review Comments

Review code comments for quality, then offer to fix issues found.

Flags

• --staged - git staged files
• --changed - git unstaged changes
• --all - entire codebase (parallel agents)
• --fix - auto-apply fixes (remove + refactor) without prompting
• Default: --staged --changed combined

Usage

/review-comments              # staged + changed (default)
/review-comments --all        # entire codebase
/review-comments --fix        # review and auto-fix

Gotchas

• Framework directives (// @ts-ignore, // eslint-disable, // noinspection) look like "what" comments but are functional. Removing them silently breaks linting or type checking.
• --staged on a clean working tree (everything already committed) reviews nothing and reports "No files to review" — which looks like a passing review. Use --all after committing.

Workflow

Step 1: Parse Flags

From $ARGUMENTS, determine scope:

Step 2: Get File List

For --staged:

git diff --cached --name-only

For --changed:

git diff --name-only

For --all:

# Use Glob to find all source files
**/*.{ts,tsx,js,jsx,py,go,java,kt,swift,rs,c,cpp,h,hpp,cs,rb,php}

Filter to only source code files (exclude node_modules, build dirs, etc.).

Step 3: Review Comments

For --staged/--changed (small file lists): Review files directly - read each file and analyze comments.

For --all (large codebase): Split files into batches and spawn parallel sub-agents:

1. Get all source files
2. Split into batches of ~20 files each
3. Spawn sub-agents in parallel using Task tool:

Subagent type: general-purpose
Prompt per batch:
---
Review comments in these files for quality issues:
{file_list}

For each file, identify:
1. "What" comments that should be "why" comments
2. Comments that could be replaced with better function/variable names
3. Obvious/redundant comments

Return findings in this format:
## {filename}
- Line {n}: "{comment}" → {suggestion}
---

1. Collect and merge results from all agents

Step 4: Report Findings

## Comment Review: {scope}

### Issues Found

#### {filename}
| Line | Current Comment | Issue | Suggestion |
|------|-----------------|-------|------------|
| {n} | "{comment}" | What not Why | {suggestion} |

### Summary
- Files reviewed: {count}
- Issues found: {count}
- Most common issue: {type}

Step 5: Offer to Fix

If --fix flag is present: Skip the prompt and directly apply Option 2 (Remove and refactor).

Otherwise, if issues were found, ask the user:

Found {n} low-quality comments. How would you like to proceed?

1. **Remove all** - Delete all flagged comments
2. **Remove and refactor** - Delete comments + extract better function/variable names where suggested
3. **Cherry-pick** - Show me each one and I'll decide
4. **Skip** - Keep the report, don't change anything

Based on user choice:

Option 1 (Remove all):

• Use Edit tool to remove each flagged comment
• Report: "Removed {n} comments from {m} files"

Option 2 (Remove and refactor):

• Remove comments
• Where suggestion was "extract to function" or "rename variable", apply that refactor
• Report changes made

Option 3 (Cherry-pick):

• Present each issue one at a time: "Line 42: // increment counter - Remove? (y/n/stop)"
• Apply user's choices

Option 4 (Skip):

• End without changes

Comment Quality Guidelines

Good Comments (WHY)

• Explain business logic decisions
• Document performance considerations
• Clarify non-obvious algorithms
• Explain workarounds or edge cases
• Document assumptions or constraints

Bad Comments (WHAT)

• Describe what the code is doing syntactically
• Restate variable names or method calls
• Explain obvious operations
• Add noise without value

Better Alternatives

Instead of "what" comments, prefer:

• Extracting functions with descriptive names
• Using meaningful variable names
• Writing self-documenting code

Examples

Bad Comment

# Loop through users and check if active
for user in users:
    if user.is_active:
        active_users.append(user)

Good Refactor

active_users = [u for u in users if u.is_active]
# Or extract to: active_users = filter_active_users(users)

Good Comment (when needed)

# Use binary search here because users list is sorted and can exceed 100k items
index = bisect.bisect_left(sorted_users, target_user)

Troubleshooting

False positives on license headers or necessary comments

Solution: License headers and legal notices are required -- skip them during review. If the tool flags doc comments (JSDoc, docstrings) or regulatory annotations, those are intentional and should be excluded from cleanup.

Comments in unfamiliar language or framework

Solution: Focus on structural patterns (e.g., comments restating the next line of code are low-quality in any language). For framework-specific annotations or directives you do not recognize, leave them in place and flag them for manual review rather than removing them.

Notes

• For --all on large codebases, parallel agents significantly speed up review
• Focus on actionable suggestions, not nitpicks
• If no files match the scope, report "No files to review"