zhongwei/gh-bejranonda-llm-autonomous-agent-plugin-for-claude

Files

Zhongwei Li c5931553a6 Initial commit

2025-11-29 18:00:50 +08:00

14 KiB

Raw Permalink Blame History

name, description, delegates-to

name	description	delegates-to
dev:pr-review	CodeRabbit-style PR review with security scanning, test coverage, and one-click fixes	autonomous-agent:pr-reviewer

Pull Request Review Command

Execute a comprehensive CodeRabbit-style review of a pull request with automated analysis, security scanning, and one-click fixes.

Usage

/dev:pr-review [PR_NUMBER|BRANCH_NAME]

Examples:

/dev:pr-review 123              # Review PR #123
/dev:pr-review feature/auth     # Review branch against main
/dev:pr-review                  # Review current branch changes

Workflow

1. Initialize Review

Detect PR context (number, branch, or current changes)
Fetch PR metadata (title, author, description)
Extract git diff and commit history

2. Delegate to PR Reviewer Agent

Execute comprehensive review via pr-reviewer agent:

const review_result = await delegate_to_pr_reviewer({
    pr_number: pr_number,
    pr_data: {
        title: pr_title,
        author: pr_author,
        description: pr_description,
        files: changed_files,
        diff: full_diff,
        commits: commit_history
    }
});

3. Analysis Pipeline

The PR reviewer agent executes:

A. Summary Generation (5-10s):

Change categorization (features, bug fixes, refactoring, etc.)
Files changed count and line statistics
Complexity score calculation

B. Line-by-Line Analysis (30-60s):

Code quality issues (naming, duplication, complexity)
Best practice violations (SOLID, DRY, error handling)
Performance concerns (N+1 queries, inefficient algorithms)
Type annotations and documentation

C. Security Scan (20-40s via security-auditor):

OWASP Top 10 vulnerability detection
Input validation checks
Authentication/authorization review
Secrets exposure detection
Dependency vulnerability scan

D. Test Coverage Analysis (15-30s):

Calculate coverage for changed lines
Identify untested functions
Generate test suggestions
Coverage delta calculation

E. Automated Fix Generation (10-20s):

Generate one-click fixes for auto-fixable issues
Provide suggestions with explanations
Calculate confidence scores

F. Risk Assessment (5-10s):

Calculate weighted risk score
Identify risk factors (size, complexity, critical files)
Generate recommendations

G. Related PR Detection (5-10s):

Find PRs touching same files
Detect similar changes
Identify dependencies

4. Report Generation

Generate comprehensive review report:

# Pull Request Review: #{PR_NUMBER}

## 📊 Summary
**Risk Level**: {RISK_LEVEL} ({RISK_SCORE}/100)
Files: {COUNT} | +{ADDITIONS} -{DELETIONS} | Complexity: {SCORE}/100

## 🔒 Security ({VULN_COUNT} issues)
🔴 Critical: {COUNT} | 🟠 High: {COUNT} | 🟡 Medium: {COUNT}

## 📈 Test Coverage
{COVERAGE}% ({DELTA > 0 ? '+' : ''}{DELTA}%) | Untested: {COUNT}

## 💡 Code Review ({ISSUE_COUNT} issues)
{DETAILED_REVIEWS_BY_FILE}

## ⚡ Performance ({ISSUE_COUNT} concerns)
{PERFORMANCE_ISSUES}

## 🎯 Recommendations
### Critical ({COUNT})
### Suggested ({COUNT})
### Nice to Have ({COUNT})

## ✅ Approval Checklist
- [ ] All critical issues resolved
- [ ] Test coverage adequate
- [ ] No new vulnerabilities
- [ ] Performance acceptable

5. Interactive Fix Application

Provide one-click fix application:

# Auto-fixable issues presented with "Apply Fix" option
# User can select fixes to apply
# System applies fixes and creates commit

Skills Integration

This command leverages:

ast-analyzer:

Deep code structure analysis
Complexity calculation
Impact analysis

security-patterns:

Vulnerability detection patterns
Secure coding guidelines

contextual-pattern-learning:

Find similar successful PRs
Learn review patterns
Improve accuracy over time

code-analysis:

Code quality metrics
Best practice validation

Output Format

Terminal Output (Tier 1: Concise Summary)

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
  PR REVIEW COMPLETE: #{PR_NUMBER}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

📊 Overview
   Risk Level: {RISK_LEVEL} ({RISK_SCORE}/100)
   Files: {COUNT} | +{ADDITIONS} -{DELETIONS}
   Complexity: {SCORE}/100

🔒 Security Analysis
   🔴 Critical: {COUNT} | 🟠 High: {COUNT} | 🟡 Medium: {COUNT}
   Total New Vulnerabilities: {COUNT}

📈 Test Coverage
   Coverage: {COVERAGE}% ({DELTA > 0 ? '+' : ''}{DELTA}%)
   Untested Functions: {COUNT}

💡 Top 3 Issues
   1. {SEVERITY} - {FILE}:{LINE} - {ISSUE}
   2. {SEVERITY} - {FILE}:{LINE} - {ISSUE}
   3. {SEVERITY} - {FILE}:{LINE} - {ISSUE}

🎯 Top 3 Recommendations
   1. {CRITICAL_RECOMMENDATION}
   2. {SUGGESTED_IMPROVEMENT}
   3. {NICE_TO_HAVE}

✅ Auto-fixable Issues: {COUNT}/{TOTAL}

📄 Detailed Report: .data/reports/pr-review/pr-{NUMBER}-{DATE}.md

⏱️  Review completed in {DURATION}
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

Detailed Report (Tier 2: Comprehensive File)

Saved to: .data/reports/pr-review/pr-{NUMBER}-{YYYY-MM-DD}.md

Full Report Structure:

# Pull Request Review: #{PR_NUMBER}
**Generated**: {TIMESTAMP}
**Review Time**: {DURATION}
**Reviewer**: Autonomous PR Review Agent v1.0
---

## Table of Contents
1. [Summary](#summary)
2. [Security Analysis](#security-analysis)
3. [Test Coverage](#test-coverage)
4. [Code Review](#code-review)
5. [Performance Analysis](#performance-analysis)
6. [Recommendations](#recommendations)
7. [Related PRs](#related-prs)
8. [Approval Checklist](#approval-checklist)

---

## Summary

**Title**: {PR_TITLE}
**Author**: {AUTHOR}
**Status**: {STATUS}
**Risk Level**: {RISK_LEVEL} ({RISK_SCORE}/100)

### Changes Overview
- **Files Changed**: {COUNT}
- **Lines Added**: +{ADDITIONS}
- **Lines Removed**: -{DELETIONS}
- **Complexity Score**: {SCORE}/100

### Change Categories
- ✨ **Features**: {COUNT} files
  - {FILE_LIST}
- 🐛 **Bug Fixes**: {COUNT} files
  - {FILE_LIST}
- ♻️  **Refactoring**: {COUNT} files
  - {FILE_LIST}
- 📝 **Documentation**: {COUNT} files
  - {FILE_LIST}
- ✅ **Tests**: {COUNT} files
  - {FILE_LIST}

### Risk Factors
| Factor | Score | Weight | Impact |
|--------|-------|--------|--------|
| Size | {SCORE}/100 | 20% | {IMPACT} |
| Complexity | {SCORE}/100 | 25% | {IMPACT} |
| Test Coverage | {SCORE}/100 | 25% | {IMPACT} |
| Critical Files | {SCORE}/100 | 20% | {IMPACT} |
| Security | {SCORE}/100 | 10% | {IMPACT} |

---

## Security Analysis

**New Vulnerabilities Detected**: {COUNT}

### Critical Issues (🔴)

#### {VULN_TITLE_1}
- **File**: `{FILE_PATH}`
- **Line**: {LINE_NUMBER}
- **Severity**: CRITICAL
- **CWE**: CWE-{NUMBER} - {CWE_NAME}
- **OWASP**: {OWASP_CATEGORY}

**Vulnerable Code**:
```{LANGUAGE}
{VULNERABLE_CODE}

Description: {DETAILED_DESCRIPTION}

Remediation:

{FIXED_CODE}

Explanation: {EXPLANATION}

Auto-fixable: {YES/NO}

[Apply Fix] (One-click button)

High Issues (🟠)

{SIMILAR_STRUCTURE}

Medium Issues (🟡)

{SIMILAR_STRUCTURE}

Low Issues (⚪)

{SIMILAR_STRUCTURE}

Test Coverage

Overall Coverage: {COVERAGE}% ({DELTA > 0 ? '+' : ''}{DELTA}%)

Changed Lines Coverage: {CHANGED_LINES_COV}%
Untested Functions: {COUNT}

Coverage by File

File	Before	After	Delta	Untested Functions
{FILE}	{BEFORE}%	{AFTER}%	{DELTA}%	{COUNT}

Untested Functions

{FILE_PATH}

{FUNCTION_NAME} (line {LINE})
{FUNCTION_NAME} (line {LINE})

Suggested Test:

{SUGGESTED_TEST_CODE}

Code Review

{FILE_PATH_1}

Line {LINE}: {ISSUE_TITLE}

Severity: {CRITICAL/HIGH/MEDIUM/LOW} Category: {CODE_QUALITY/BEST_PRACTICE/PERFORMANCE}

Original Code:

{ORIGINAL_CODE}

Issue: {DETAILED_ISSUE_DESCRIPTION}

Suggested Fix:

{SUGGESTED_CODE}

Explanation: {WHY_THIS_IS_BETTER}

Auto-fixable: {YES/NO} Confidence: {CONFIDENCE}%

[Apply Fix] (One-click button)

{FILE_PATH_2}

{SIMILAR_STRUCTURE}

Performance Analysis

Potential Performance Impact: {LOW/MEDIUM/HIGH}

N+1 Query Issues ({COUNT})

{FILE}:{LINE} - {FUNCTION_NAME}

Detected Pattern: Loop with database query inside

Current Code:

{CURRENT_CODE}

Optimized Code:

{OPTIMIZED_CODE}

Performance Improvement: {ESTIMATED_IMPROVEMENT}

Inefficient Algorithms ({COUNT})

{SIMILAR_STRUCTURE}

Missing Indexes ({COUNT})

{SIMILAR_STRUCTURE}

Large Data Operations ({COUNT})

{SIMILAR_STRUCTURE}

Recommendations

🔴 Critical Actions Required ({COUNT})

{CRITICAL_ISSUE_1}
- File: {FILE}
- Action: {SPECIFIC_ACTION}
- Impact: {IMPACT_DESCRIPTION}
{CRITICAL_ISSUE_2} {SIMILAR_STRUCTURE}

🟡 Suggested Improvements ({COUNT})

{IMPROVEMENT_1}
- File: {FILE}
- Benefit: {BENEFIT_DESCRIPTION}
- Effort: {LOW/MEDIUM/HIGH}
{IMPROVEMENT_2} {SIMILAR_STRUCTURE}

⚪ Nice to Have ({COUNT})

{NICE_TO_HAVE_1}
- File: {FILE}
- Benefit: {MINOR_BENEFIT}

PRs Touching Same Files

#{PR_NUMBER}: {TITLE}
- Author: {AUTHOR}
- Status: {STATUS}
- Overlap: {FILE_COUNT} files
- Potential Conflict: {YES/NO}

Similar PRs

#{PR_NUMBER}: {TITLE}
- Similarity: {PERCENTAGE}%
- Lessons Learned: {INSIGHTS}

Dependent PRs

#{PR_NUMBER}: {TITLE}
- Dependency Type: {BLOCKS/BLOCKED_BY}

Approval Checklist

Mandatory Requirements

All critical security issues resolved
Test coverage ≥ 70% for changed lines
No new critical vulnerabilities introduced
All tests passing
Documentation updated

Code Quality

No code quality issues with severity > MEDIUM
Best practices followed
Performance impact acceptable
No technical debt introduced

Review Sign-off

Security review complete
Performance review complete
Test coverage adequate
Code review complete

Review Metadata

Review Generated: {TIMESTAMP} Review Time: {DURATION} Auto-fixable Issues: {COUNT}/{TOTAL} Confidence Score: {AVERAGE_CONFIDENCE}%

Reviewer Agent: pr-reviewer v1.0 Security Scanner: security-auditor v1.0 AST Analyzer: ast-analyzer v1.0 Pattern Learner: contextual-pattern-learning v3.0

One-Click Fixes Available

{COUNT} issues can be fixed automatically. Apply all fixes with:

/apply-pr-fixes {PR_NUMBER}

Or apply individual fixes:

/apply-fix {ISSUE_ID}

End of Report


---

## Implementation Details

### Git Integration

```python
def fetch_pr_data(pr_identifier):
    """Fetch PR data from git or GitHub CLI."""
    if pr_identifier.isdigit():
        # Use gh CLI for PR number
        pr_data = subprocess.run(
            ["gh", "pr", "view", pr_identifier, "--json",
             "title,author,body,files,additions,deletions"],
            capture_output=True
        )
    else:
        # Use git for branch comparison
        diff = subprocess.run(
            ["git", "diff", f"origin/main...{pr_identifier}"],
            capture_output=True
        )
        commits = subprocess.run(
            ["git", "log", f"origin/main..{pr_identifier}",
             "--oneline"],
            capture_output=True
        )

    return parse_pr_data(pr_data)

Fix Application

def apply_fix(issue_id):
    """Apply automated fix for specific issue."""
    issue = load_issue(issue_id)

    if not issue.auto_fixable:
        print("Issue not auto-fixable")
        return False

    # Apply Edit tool
    Edit(
        file_path=issue.file,
        old_string=issue.original_code,
        new_string=issue.suggested_code
    )

    # Run tests to verify
    test_result = run_tests()

    if test_result.success:
        # Create commit
        git_commit(f"Fix: {issue.title}\n\nAuto-applied fix from PR review")
        return True
    else:
        # Rollback
        git_checkout(issue.file)
        return False

Learning Integration

After each PR review, the learning engine captures:

Review Patterns:
- Which issues were found in which file types
- Success rate of automated fixes
- False positive rates
Project Patterns:
- Common issue patterns in this codebase
- Team coding style preferences
- Review thoroughness preferences
Performance Metrics:
- Review time by PR size
- Issue detection accuracy
- Fix application success rate
Continuous Improvement:
- Reduce false positives over time
- Improve fix suggestion quality
- Personalize review style to team

Error Handling

try:
    review_result = comprehensive_pr_review(pr_number)
except GitError as e:
    print(f"Git error: {e.message}")
    print("Ensure you're in a git repository and PR exists")
except SecurityScanError as e:
    print(f"Security scan failed: {e.message}")
    print("Review will continue with partial results")
except Exception as e:
    print(f"Review failed: {e}")
    print("Saving partial results...")
    save_partial_review(partial_data)

Performance Expectations

PR Size	Files	Lines	Review Time
Small	1-5	<200	30-60s
Medium	6-15	200-500	1-2min
Large	16-30	500-1000	2-4min
XLarge	31+	1000+	4-8min

Follow-up Commands

After review:

/apply-pr-fixes {PR_NUMBER}     # Apply all auto-fixable issues
/apply-fix {ISSUE_ID}           # Apply specific fix
/dev:pr-review-history          # Show review history
/learn:analytics                # Review performance analytics

This command provides comprehensive, CodeRabbit-level PR review capabilities with deep integration into the autonomous learning system.

14 KiB Raw Permalink Blame History

Pull Request Review Command

Usage

Workflow

1. Initialize Review

2. Delegate to PR Reviewer Agent

3. Analysis Pipeline

4. Report Generation

5. Interactive Fix Application

Skills Integration

Output Format

Terminal Output (Tier 1: Concise Summary)

Detailed Report (Tier 2: Comprehensive File)

High Issues (🟠)

Medium Issues (🟡)

Low Issues (⚪)

Test Coverage

Coverage by File

Untested Functions

{FILE_PATH}

Code Review

{FILE_PATH_1}

Line {LINE}: {ISSUE_TITLE}

{FILE_PATH_2}

Performance Analysis

N+1 Query Issues ({COUNT})

{FILE}:{LINE} - {FUNCTION_NAME}

Inefficient Algorithms ({COUNT})

Missing Indexes ({COUNT})

Large Data Operations ({COUNT})

Recommendations

🔴 Critical Actions Required ({COUNT})

🟡 Suggested Improvements ({COUNT})

⚪ Nice to Have ({COUNT})

Related PRs

PRs Touching Same Files

Similar PRs

Dependent PRs

Approval Checklist

Mandatory Requirements

Code Quality

Review Sign-off

Review Metadata

One-Click Fixes Available

Fix Application

Learning Integration

Error Handling

Performance Expectations

Follow-up Commands

14 KiB

Raw Permalink Blame History