antipattern-detection-bundle

# Anti-Pattern Detection Bundle Agent uses this skill to identify, challenge, and remediate anti-patterns before creation or during refactoring across performance, code quality, design, and architectural dimensions. ## When to Use | Condition | Use | |---|---| | Agent reviews code or design for quality issues | Use this skill | | Developer proposes new pattern or architecture | Use this skill | | Refactoring work targets existing codebase | Use this skill | | Performance or scalability concerns exist | Use this skill | | Code review identifies potential anti-pattern | Use this skill | ## When Not to Use | Condition | Alternative | |---|---| | Specific Azure anti-pattern already identified | Use `performance-antipatterns` directly | | Code smell analysis only | Use `code-smell-detection` directly | | Design debate without anti-pattern context | Use `architectural-decision-debate` directly | ## Required Inputs | Input | Required | Notes | |---|---|---| | Code, design, or architecture description | Yes | Context for anti-pattern detection | | Scope | Yes | New code, existing code, or design proposal | | Performance or quality concerns | No | Specific symptoms guide detection | ## Anti-Pattern Detection Framework Agent applies layered anti-pattern detection by delegating to specialized skills: ### Detection Layer Routing | Layer | Skill | Token Cost | Coverage | |---|---|---|---| | Performance & Scalability | `performance-antipatterns` | ~1200 | 10 Azure anti-patterns with references | | Code Quality | `code-smell-detection` | ~1100 | 10 code smells with references | | Structural Design | `design-smell-detection` | ~900 | 7 design smells | | Architectural Decisions | `architectural-decision-debate` | ~3200 | 3-position debate framework | **Total bundle token cost:** ~6400 tokens (excluding reference files) ### When to Use Sub-Skills Directly - **Known anti-pattern category:** Use specific skill directly - **Broad quality review:** Use this bundle to scan all dimensions - **Decision debate needed:** Use `architectural-decision-debate` ## Quick Reference **Azure Performance Anti-Patterns:** - Busy Database, Busy Front End, Chatty I/O, Extraneous Fetching - Improper Instantiation, Monolithic Persistence, No Caching - Noisy Neighbor, Retry Storm, Synchronous I/O **Code Smells:** - Duplicated Code, Long Method, Large Class, Too Many Parameters - Data Clump, Shotgun Surgery, Feature Envy, Magic Numbers - Primitive Obsession, Refused Bequest **Design Smells:** - Cyclic Dependencies, Tight Coupling, Insufficient Modularization - Broken Hierarchy, Missing Abstraction, Unstable Dependencies **Detailed guidance:** Each sub-skill contains `references/` directory with specific anti-pattern remediation examples. ## Workflow 1. Agent identifies scope (new code, existing code, design proposal). 2. Agent routes to appropriate detection skill(s): - Performance concerns → `performance-antipatterns` - Code quality concerns → `code-smell-detection` - Structural concerns → `design-smell-detection` - Decision debate → `architectural-decision-debate` 3. Agent consolidates findings from sub-skills. 4. Agent reports detected anti-patterns with severity, evidence, and impact. 5. Agent proposes refactoring alternatives or design changes. 6. Agent validates proposed changes eliminate anti-pattern without introducing new issues. Test: Run static analysis, build, and tests after remediation. Pass: Zero new defects. Anti-pattern eliminated. All tests pass. ## Detection Rules Matrix | Rule | Dimension | Detection Pattern | Remediation | |---|---|---|---| | ANTI-001 | Azure | Database performs business logic | Move logic to application tier | | ANTI-002 | Azure | Request threads perform CPU-intensive work | Use background processing | | ANTI-003 | Azure | Multiple small HTTP calls in loop | Batch requests or use bulk operations | | ANTI-004 | Azure | Fetching entire dataset when subset needed | Add filtering, projection, pagination | | ANTI-005 | Azure | Creating DbContext or HttpClient per request | Use DI with correct lifetime | | ANTI-006 | Azure | Single database for OLTP and analytics | Separate read and write stores | | ANTI-007 | Azure | No caching for frequently read static data | Add caching layer with expiration | | ANTI-008 | Azure | Synchronous I/O in async method | Use true async I/O operations | | ANTI-009 | Code | Duplicated logic blocks >10 lines | Extract to shared method or class | | ANTI-010 | Code | Method exceeds 50 lines | Split into smaller focused methods | | ANTI-011 | Code | Class has >10 public methods | Apply Single Responsibility Principle | | ANTI-012 | Code | Method has >5 parameters | Introduce parameter object or builder | | ANTI-013 | Code | Magic numbers without explanation | Replace with named constants | | ANTI-014 | Code | Primitive types for domain concepts | Create value objects | | ANTI-015 | Design | Circular namespace or project references | Break cycle with abstraction or event | | ANTI-016 | Design | Concrete class depends on concrete class | Introduce interface or abstraction | | ANTI-017 | Arch | No clear layer separation | Define and enforce architectural boundaries | | ANTI-018 | Arch | Dead code remains uncommented | Remove or document as intentional | ## Verification Matrix | Test | Run | Pass | |---|---|---| | Static analysis | `dotnet format analyzers` or static analyzer | Zero new violations | | Code metrics | Analyze complexity, maintainability index | Metrics improve or stay within thresholds | | Build verification | `dotnet build` | Zero compile errors | | Test verification | `dotnet test` | All tests pass | | Performance baseline | Compare before/after performance | Performance improves or remains stable | ## Prevention Guidance Agent provides anti-pattern prevention when generating new code: - **Before adding database logic:** Consider if logic belongs in application tier - **Before synchronous I/O:** Use async/await throughout - **Before caching bypass:** Check if data is frequently read and static - **Before concrete dependency:** Introduce abstraction for testability - **Before inline logic duplication:** Extract to reusable component - **Before large method:** Split into focused single-purpose methods ## Remediation Workflow When anti-pattern detected: 1. Agent classifies anti-pattern by dimension and severity. 2. Agent collects evidence (code snippets, metrics, references). 3. Agent explains why pattern is problematic (performance, maintainability, testability). 4. Agent proposes specific refactoring with before/after examples. 5. Agent estimates impact (lines changed, test changes, risk). 6. Agent validates remediation eliminates issue. ## Integration with Other Skills | Anti-Pattern Type | Primary Skill | Secondary Skills | |---|---|---| | Azure performance anti-patterns | `performance-antipatterns` | `analyzing-dotnet-performance`, `optimizing-ef-core-queries` | | Code smells | `code-smell-detection` | `refactor`, `dotnet-code-quality-standards` | | Design smells | `design-smell-detection` | `dotnet-architectural-layers`, `dependency-injection-patterns` | | Architectural decisions | `architectural-decision-debate` | `technology-selection`, `api-gateway-patterns` | ## Output Format Agent reports detected anti-patterns in structured format: ``` Anti-Pattern Detection Report ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Found: 3 anti-patterns 1. [HIGH] Busy Database Location: CustomerRepository.cs, lines 45-78 Evidence: Stored procedure performs complex business calculations Impact: Database CPU contention under load Remediation: Move calculation logic to CustomerEngine service Effort: ~4 hours (extract logic, add unit tests) 2. [MEDIUM] Duplicated Code Location: OrderService.cs (lines 120-135), InvoiceService.cs (lines 88-103) Evidence: Identical tax calculation logic in both classes Impact: Maintenance burden, inconsistency risk Remediation: Extract to TaxCalculator utility class Effort: ~1 hour 3. [LOW] Magic Numbers Location: DiscountService.cs, lines 23, 45, 67 Evidence: Unexplained numeric literals (0.15, 0.20, 0.25) Impact: Reduced readability Remediation: Replace with named constants (StandardDiscount, PremiumDiscount, VipDiscount) Effort: ~30 minutes ``` ## Verification Checklist Agent verifies: - [ ] All four detection dimensions scanned - [ ] Detected anti-patterns classified by severity - [ ] Evidence collected for each detection - [ ] Remediation proposed with effort estimate - [ ] No new anti-patterns introduced by remediation - [ ] Performance and quality metrics remain stable or improve