/full-review

Comprehensive Code Review Orchestrator

CRITICAL BEHAVIORAL RULES

You MUST follow these rules exactly. Violating any of them is a failure.

1. Execute phases in order. Do NOT skip ahead, reorder, or merge phases.
2. Write output files. Each phase MUST produce its output file in .full-review/ before the next phase begins. Read from prior phase files -- do NOT rely on context window memory.
3. Stop at checkpoints. When you reach a PHASE CHECKPOINT, you MUST stop and wait for explicit user approval before continuing. Use the AskUserQuestion tool with clear options.
4. Halt on failure. If any step fails (agent error, missing files, access issues), STOP immediately. Present the error and ask the user how to proceed. Do NOT silently continue.
5. Use only local agents. All subagent_type references use agents bundled with this plugin or general-purpose. No cross-plugin dependencies.
6. Never enter plan mode autonomously. Do NOT use EnterPlanMode. This command IS the plan -- execute it.

Pre-flight Checks

Before starting, perform these checks:

1. Check for existing session

Check if .full-review/state.json exists:

• If it exists and status is "in_progress": Read it, display the current phase, and ask the user:

  Found an in-progress review session:
  Target: [target from state]
  Current phase: [phase from state]
  
  1. Resume from where we left off
  2. Start fresh (archives existing session)
  

• If it exists and status is "complete": Ask whether to archive and start fresh.

2. Initialize state

Create .full-review/ directory and state.json:

{
  "target": "$ARGUMENTS",
  "status": "in_progress",
  "flags": {
    "security_focus": false,
    "performance_critical": false,
    "strict_mode": false,
    "framework": null
  },
  "current_step": 1,
  "current_phase": 1,
  "completed_steps": [],
  "files_created": [],
  "started_at": "ISO_TIMESTAMP",
  "last_updated": "ISO_TIMESTAMP"
}

Parse $ARGUMENTS for --security-focus, --performance-critical, --strict-mode, and --framework flags. Update the flags object accordingly.

3. Identify review target

Determine what code to review from $ARGUMENTS:

• If a file/directory path is given, verify it exists
• If a description is given (e.g., "recent changes", "authentication module"), identify the relevant files
• List the files that will be reviewed and confirm with the user

Output file: .full-review/00-scope.md

# Review Scope

## Target

[Description of what is being reviewed]

## Files

[List of files/directories included in the review]

## Flags

- Security Focus: [yes/no]
- Performance Critical: [yes/no]
- Strict Mode: [yes/no]
- Framework: [name or auto-detected]

## Review Phases

1. Code Quality & Architecture
2. Security & Performance
3. Testing & Documentation
4. Best Practices & Standards
5. Consolidated Report

Update state.json: add "00-scope.md" to files_created, add step 0 to completed_steps.

---

Phase 1: Code Quality & Architecture Review (Steps 1A-1B)

Run both agents in parallel using multiple Task tool calls in a single response.

Step 1A: Code Quality Analysis

Task:
  subagent_type: "code-reviewer"
  description: "Code quality analysis for $ARGUMENTS"
  prompt: |
    Perform a comprehensive code quality review.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Instructions
    Analyze the target code for:
    1. **Code complexity**: Cyclomatic complexity, cognitive complexity, deeply nested logic
    2. **Maintainability**: Naming conventions, function/method length, class cohesion
    3. **Code duplication**: Copy-pasted logic, missed abstraction opportunities
    4. **Clean Code principles**: SOLID violations, code smells, anti-patterns
    5. **Technical debt**: Areas that will become increasingly costly to change
    6. **Error handling**: Missing error handling, swallowed exceptions, unclear error messages

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - File and line location
    - Description of the issue
    - Specific fix recommendation with code example

    Write your findings as a structured markdown document.

Step 1B: Architecture & Design Review

Task:
  subagent_type: "architect-review"
  description: "Architecture review for $ARGUMENTS"
  prompt: |
    Review the architectural design and structural integrity of the target code.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Instructions
    Evaluate the code for:
    1. **Component boundaries**: Proper separation of concerns, module cohesion
    2. **Dependency management**: Circular dependencies, inappropriate coupling, dependency direction
    3. **API design**: Endpoint design, request/response schemas, error contracts, versioning
    4. **Data model**: Schema design, relationships, data access patterns
    5. **Design patterns**: Appropriate use of patterns, missing abstractions, over-engineering
    6. **Architectural consistency**: Does the code follow the project's established patterns?

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - Architectural impact assessment
    - Specific improvement recommendation

    Write your findings as a structured markdown document.

After both complete, consolidate into .full-review/01-quality-architecture.md:

# Phase 1: Code Quality & Architecture Review

## Code Quality Findings

[Summary from 1A, organized by severity]

## Architecture Findings

[Summary from 1B, organized by severity]

## Critical Issues for Phase 2 Context

[List any findings that should inform security or performance review]

Update state.json: set current_step to 2, current_phase to 2, add steps 1A and 1B to completed_steps.

---

Phase 2: Security & Performance Review (Steps 2A-2B)

Read .full-review/01-quality-architecture.md for context from Phase 1.

Run both agents in parallel using multiple Task tool calls in a single response.

Step 2A: Security Vulnerability Assessment

Task:
  subagent_type: "security-auditor"
  description: "Security audit for $ARGUMENTS"
  prompt: |
    Execute a comprehensive security audit on the target code.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Phase 1 Context
    [Insert contents of .full-review/01-quality-architecture.md -- focus on the "Critical Issues for Phase 2 Context" section]

    ## Instructions
    Analyze for:
    1. **OWASP Top 10**: Injection, broken auth, sensitive data exposure, XXE, broken access control, misconfig, XSS, insecure deserialization, vulnerable components, insufficient logging
    2. **Input validation**: Missing sanitization, unvalidated redirects, path traversal
    3. **Authentication/authorization**: Flawed auth logic, privilege escalation, session management
    4. **Cryptographic issues**: Weak algorithms, hardcoded secrets, improper key management
    5. **Dependency vulnerabilities**: Known CVEs in dependencies, outdated packages
    6. **Configuration security**: Debug mode, verbose errors, permissive CORS, missing security headers

    For each finding, provide:
    - Severity (Critical / High / Medium / Low) with CVSS score if applicable
    - CWE reference where applicable
    - File and line location
    - Proof of concept or attack scenario
    - Specific remediation steps with code example

    Write your findings as a structured markdown document.

Step 2B: Performance & Scalability Analysis

Task:
  subagent_type: "general-purpose"
  description: "Performance analysis for $ARGUMENTS"
  prompt: |
    You are a performance engineer. Conduct a performance and scalability analysis of the target code.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Phase 1 Context
    [Insert contents of .full-review/01-quality-architecture.md -- focus on the "Critical Issues for Phase 2 Context" section]

    ## Instructions
    Analyze for:
    1. **Database performance**: N+1 queries, missing indexes, unoptimized queries, connection pool sizing
    2. **Memory management**: Memory leaks, unbounded collections, large object allocation
    3. **Caching opportunities**: Missing caching, stale cache risks, cache invalidation issues
    4. **I/O bottlenecks**: Synchronous blocking calls, missing pagination, large payloads
    5. **Concurrency issues**: Race conditions, deadlocks, thread safety
    6. **Frontend performance**: Bundle size, render performance, unnecessary re-renders, missing lazy loading
    7. **Scalability concerns**: Horizontal scaling barriers, stateful components, single points of failure

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - Estimated performance impact
    - Specific optimization recommendation with code example

    Write your findings as a structured markdown document.

After both complete, consolidate into .full-review/02-security-performance.md:

# Phase 2: Security & Performance Review

## Security Findings

[Summary from 2A, organized by severity]

## Performance Findings

[Summary from 2B, organized by severity]

## Critical Issues for Phase 3 Context

[List findings that affect testing or documentation requirements]

Update state.json: set current_step to "checkpoint-1", add steps 2A and 2B to completed_steps.

---

PHASE CHECKPOINT 1 -- User Approval Required

Display a summary of findings from Phase 1 and Phase 2 and ask:

Phases 1-2 complete: Code Quality, Architecture, Security, and Performance reviews done.

Summary:
- Code Quality: [X critical, Y high, Z medium findings]
- Architecture: [X critical, Y high, Z medium findings]
- Security: [X critical, Y high, Z medium findings]
- Performance: [X critical, Y high, Z medium findings]

Please review:
- .full-review/01-quality-architecture.md
- .full-review/02-security-performance.md

1. Continue -- proceed to Testing & Documentation review
2. Fix critical issues first -- I'll address findings before continuing
3. Pause -- save progress and stop here

If --strict-mode flag is set and there are Critical findings, recommend option 2.

Do NOT proceed to Phase 3 until the user approves.

---

Phase 3: Testing & Documentation Review (Steps 3A-3B)

Read .full-review/01-quality-architecture.md and .full-review/02-security-performance.md for context.

Run both agents in parallel using multiple Task tool calls in a single response.

Step 3A: Test Coverage & Quality Analysis

Task:
  subagent_type: "general-purpose"
  description: "Test coverage analysis for $ARGUMENTS"
  prompt: |
    You are a test automation engineer. Evaluate the testing strategy and coverage for the target code.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Prior Phase Context
    [Insert security and performance findings from .full-review/02-security-performance.md that affect testing requirements]

    ## Instructions
    Analyze:
    1. **Test coverage**: Which code paths have tests? Which critical paths are untested?
    2. **Test quality**: Are tests testing behavior or implementation? Assertion quality?
    3. **Test pyramid adherence**: Unit vs integration vs E2E test ratio
    4. **Edge cases**: Are boundary conditions, error paths, and concurrent scenarios tested?
    5. **Test maintainability**: Test isolation, mock usage, flaky test indicators
    6. **Security test gaps**: Are security-critical paths tested? Auth, input validation, etc.
    7. **Performance test gaps**: Are performance-critical paths tested? Load testing?

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - What is untested or poorly tested
    - Specific test recommendations with example test code

    Write your findings as a structured markdown document.

Step 3B: Documentation & API Review

Task:
  subagent_type: "general-purpose"
  description: "Documentation review for $ARGUMENTS"
  prompt: |
    You are a technical documentation architect. Review documentation completeness and accuracy.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Prior Phase Context
    [Insert key findings from .full-review/01-quality-architecture.md and .full-review/02-security-performance.md]

    ## Instructions
    Evaluate:
    1. **Inline documentation**: Are complex algorithms and business logic explained?
    2. **API documentation**: Are endpoints documented with examples? Request/response schemas?
    3. **Architecture documentation**: ADRs, system diagrams, component documentation
    4. **README completeness**: Setup instructions, development workflow, deployment guide
    5. **Accuracy**: Does documentation match the actual implementation?
    6. **Changelog/migration guides**: Are breaking changes documented?

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - What is missing or inaccurate
    - Specific documentation recommendation

    Write your findings as a structured markdown document.

After both complete, consolidate into .full-review/03-testing-documentation.md:

# Phase 3: Testing & Documentation Review

## Test Coverage Findings

[Summary from 3A, organized by severity]

## Documentation Findings

[Summary from 3B, organized by severity]

Update state.json: set current_step to 4, current_phase to 4, add steps 3A and 3B to completed_steps.

---

Phase 4: Best Practices & Standards (Steps 4A-4B)

Read all previous .full-review/*.md files for full context.

Run both agents in parallel using multiple Task tool calls in a single response.

Step 4A: Framework & Language Best Practices

Task:
  subagent_type: "general-purpose"
  description: "Framework best practices review for $ARGUMENTS"
  prompt: |
    You are an expert in modern framework and language best practices. Verify adherence to current standards.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## All Prior Findings
    [Insert a concise summary of critical/high findings from all prior phases]

    ## Instructions
    Check for:
    1. **Language idioms**: Is the code idiomatic for its language? Modern syntax and features?
    2. **Framework patterns**: Does it follow the framework's recommended patterns? (e.g., React hooks, Django views, Spring beans)
    3. **Deprecated APIs**: Are any deprecated functions/libraries/patterns used?
    4. **Modernization opportunities**: Where could modern language/framework features simplify code?
    5. **Package management**: Are dependencies up-to-date? Unnecessary dependencies?
    6. **Build configuration**: Is the build optimized? Development vs production settings?

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - Current pattern vs recommended pattern
    - Migration/fix recommendation with code example

    Write your findings as a structured markdown document.

Step 4B: CI/CD & DevOps Practices Review

Task:
  subagent_type: "general-purpose"
  description: "CI/CD and DevOps practices review for $ARGUMENTS"
  prompt: |
    You are a DevOps engineer. Review CI/CD pipeline and operational practices.

    ## Review Scope
    [Insert contents of .full-review/00-scope.md]

    ## Critical Issues from Prior Phases
    [Insert critical/high findings from all prior phases that impact deployment or operations]

    ## Instructions
    Evaluate:
    1. **CI/CD pipeline**: Build automation, test gates, deployment stages, security scanning
    2. **Deployment strategy**: Blue-green, canary, rollback capabilities
    3. **Infrastructure as Code**: Are infrastructure configs version-controlled and reviewed?
    4. **Monitoring & observability**: Logging, metrics, alerting, dashboards
    5. **Incident response**: Runbooks, on-call procedures, rollback plans
    6. **Environment management**: Config separation, secret management, parity between environments

    For each finding, provide:
    - Severity (Critical / High / Medium / Low)
    - Operational risk assessment
    - Specific improvement recommendation

    Write your findings as a structured markdown document.

After both complete, consolidate into .full-review/04-best-practices.md:

# Phase 4: Best Practices & Standards

## Framework & Language Findings

[Summary from 4A, organized by severity]

## CI/CD & DevOps Findings

[Summary from 4B, organized by severity]

Update state.json: set current_step to 5, current_phase to 5, add steps 4A and 4B to completed_steps.

---

Phase 5: Consolidated Report (Step 5)

Read all .full-review/*.md files. Generate the final consolidated report.

Output file: .full-review/05-final-report.md

# Comprehensive Code Review Report

## Review Target

[From 00-scope.md]

## Executive Summary

[2-3 sentence overview of overall code health and key concerns]

## Findings by Priority

### Critical Issues (P0 -- Must Fix Immediately)

[All Critical findings from all phases, with source phase reference]

- Security vulnerabilities with CVSS > 7.0
- Data loss or corruption risks
- Authentication/authorization bypasses
- Production stability threats

### High Priority (P1 -- Fix Before Next Release)

[All High findings from all phases]

- Performance bottlenecks impacting user experience
- Missing critical test coverage
- Architectural anti-patterns causing technical debt
- Outdated dependencies with known vulnerabilities

### Medium Priority (P2 -- Plan for Next Sprint)

[All Medium findings from all phases]

- Non-critical performance optimizations
- Documentation gaps
- Code refactoring opportunities
- Test quality improvements

### Low Priority (P3 -- Track in Backlog)

[All Low findings from all phases]

- Style guide violations
- Minor code smell issues
- Nice-to-have improvements

## Findings by Category

- **Code Quality**: [count] findings ([breakdown by severity])
- **Architecture**: [count] findings ([breakdown by severity])
- **Security**: [count] findings ([breakdown by severity])
- **Performance**: [count] findings ([breakdown by severity])
- **Testing**: [count] findings ([breakdown by severity])
- **Documentation**: [count] findings ([breakdown by severity])
- **Best Practices**: [count] findings ([breakdown by severity])
- **CI/CD & DevOps**: [count] findings ([breakdown by severity])

## Recommended Action Plan

1. [Ordered list of recommended actions, starting with critical/high items]
2. [Group related fixes where possible]
3. [Estimate relative effort: small/medium/large]

## Review Metadata

- Review date: [timestamp]
- Phases completed: [list]
- Flags applied: [list active flags]

Update state.json: set status to "complete", last_updated to current timestamp.

---

Completion

Present the final summary:

Comprehensive code review complete for: $ARGUMENTS

## Review Output Files
- Scope: .full-review/00-scope.md
- Quality & Architecture: .full-review/01-quality-architecture.md
- Security & Performance: .full-review/02-security-performance.md
- Testing & Documentation: .full-review/03-testing-documentation.md
- Best Practices: .full-review/04-best-practices.md
- Final Report: .full-review/05-final-report.md

## Summary
- Total findings: [count]
- Critical: [X] | High: [Y] | Medium: [Z] | Low: [W]

## Next Steps
1. Review the full report at .full-review/05-final-report.md
2. Address Critical (P0) issues immediately
3. Plan High (P1) fixes for current sprint
4. Add Medium (P2) and Low (P3) items to backlog