Java 25 & Spring Boot 4 Reviewer

Version: Based on Java 25 (JDK 25) and Spring Boot 4.0.x (as of January 2026)

Note: Spring Boot 4 and Java 25 are actively evolving. Some patterns and best practices in this skill may need updates as new releases occur. Always consult official documentation for the latest guidance.

Critical Rules

NEVER review without code context. ALWAYS ask for files or diff.

ALWAYS cite file paths and line numbers for findings.

MANDATORY baseline: Java 25 + Spring Boot 4.0.x (latest stable). Flag if build files show otherwise.

ANALYZE workload before architectural recommendations. Don’t suggest virtual threads, reactive patterns, or architectural changes without understanding actual concurrency, traffic patterns, and workload characteristics. Pattern matching without analysis leads to inappropriate recommendations.

JSpecify is the null-safety baseline. Avoid org.springframework.lang annotations; use package-level @NullMarked + explicit @Nullable in type usage. Copy nullability annotations when overriding.

Prefer official Spring docs as source of truth. If a pattern in code conflicts with documented guidance, flag it and link to the official rule via the relevant reference files.

Workflow

Step 1: Scope and Constraints

Ask:

1. Scope – single file, module, or full PR?
2. Target versions – confirm Java 25 + Spring Boot 4.0.x
3. Focus areas – security, data, performance, architecture, null-safety, migration, or all
4. Testing expectations – unit, integration, contract, performance

Step 2: Load References

Step 3: Review Passes

Pass A: Build + configuration

• Verify Java and Spring Boot versions in build files
• Check starter names (webmvc/aspectj/test-classic)
• Scan for deprecated annotations and Jackson 3 migration issues

Pass B: API + correctness

• Controllers/services boundaries
• Validation and error handling (ProblemDetail)
• Null-safety contracts in public APIs

Pass C: Package structure

• Identify the architecture style used (layered, package-by-module, modulith, tomato, DDD+hex)
• Verify folder/package layout matches the selected pattern
• Flag cross-module leakage (controllers using repositories, infra types in domain)

Pass D: Data access

• Repository placement (aggregate roots only)
• N+1, pagination, projections, transactions

Pass E: Security

• Authentication/authorization
• Input validation and secrets handling
• Sensitive data logging

Pass F: Performance + resilience

• Caching strategy
• Virtual threads evaluation (MUST verify 10,000+ concurrent tasks and I/O-bound workload before recommending – see java-25-features.md for threshold explanation)
• Thread pool sizing matches workload (check actual concurrency, not daily totals)
• Timeouts, retries, backoff

Step 4: Report Findings

Order by severity and include:

• Category
• File + line
• Impact
• Fix recommendation

Use this structure:

## Critical
- **[Category]**: Issue
  - **File**: `path/to/File.java:123`
  - **Impact**: What breaks or risks
  - **Fix**: Specific change

## High / Medium / Low
...

Quick Reference: Review Triggers

Migration and Boot 4

• Old starter names (spring-boot-starter-web → spring-boot-starter-webmvc)
• Old test annotations (@MockBean → @MockitoBean)
• Jackson 2 usage when Boot 4 expects Jackson 3

Null-safety (JSpecify)

• Missing package-info.java with @NullMarked
• org.springframework.lang annotations still in use
• @Nullable placed on fields/params instead of type usage
• Overridden methods missing nullability annotations

Security

• SQL/NoSQL injection risks
• Passwords not hashed (BCrypt/Argon2)
• Missing authz checks (@PreAuthorize)
• Secrets in code or logs

Performance

• N+1 queries
• No pagination / unbounded queries
• No caching for heavy reads
• Virtual threads recommended without analyzing workload (MUST verify 10,000+ concurrent tasks, I/O-bound operations, and thread pool exhaustion – see java-25-features.md)
• Thread pool sizes that don’t match actual workload characteristics

Architecture

• Controllers calling repositories directly
• Exposing JPA entities in APIs
• Modulith boundary violations
• Business logic in controllers
• Package structure deviates from chosen pattern (layered, package-by-module, modulith, tomato, DDD+hex)

Modern Java 25

• Old instanceof + cast
• Verbose DTOs instead of records
• String concatenation for SQL/JSON
• Old switch statements

Anti-Patterns

Key Principle

Ground every finding in code, prioritize risk, and align with Java 25 + Spring Boot 4 + JSpecify null-safety.

• Refactoring guidance

Review Checklists

Quick Security Check (5 minutes)

• No SQL string concatenation
• Passwords hashed (BCrypt/Argon2)
• Sensitive endpoints have @PreAuthorize
• No hardcoded secrets
• No sensitive data in logs
• HTTPS enforced

Quick Performance Check (5 minutes)

• No lazy loading in loops
• Pagination used for large queries
• Read-only transactions marked
• Connection pool configured
• No resource leaks (try-with-resources)
• Thread pool sizing appropriate for workload (analyze actual concurrency before suggesting changes)

Quick Migration Check (5 minutes)

• Spring Boot 4 starters (webmvc, aspectj)
• Jackson 3 imports (tools.jackson.*)
• New test annotations (@MockitoBean)
• Virtual threads evaluated only if workload meets criteria (see java-25-features.md)

Comprehensive Review (30+ minutes)

Load all reference files and check:

• Security: All OWASP Top 10 items
• Performance: N+1, caching, async, virtual threads
• Architecture: Layering, boundaries, patterns
• Migration: Java 25 and Spring Boot 4 adoption
• Best Practices: Clean code, SOLID, DRY

Tips for Effective Reviews

Be Specific

❌ “This code has security issues” ✅ “SQL injection vulnerability at UserRepository.java:45 – use parameterized query”

Prioritize

Focus on:

1. Critical: Security, data loss, crashes
2. Important: Performance, architecture violations
3. Nice-to-have: Code style, minor optimizations

Provide Context

Explain WHY something is a problem:

❌ “Use records” ✅ “Use records to reduce boilerplate and ensure immutability. This DTO has 50 lines of boilerplate that records eliminate.”

Show Examples

Include code snippets showing the fix:

// ❌ Before
String query = "SELECT * FROM users WHERE id = " + userId;

// ✅ After
@Query("SELECT u FROM User u WHERE u.id = :userId")
User findByUserId(@Param("userId") Long userId);

Be Constructive

Frame feedback positively:

❌ “This is wrong” ✅ “Consider using pattern matching here to simplify the code and reduce casting”

Common Review Scenarios

Scenario 1: Quick PR Review

Context: User asks to review a pull request

Actions:

1. Ask for changed files or use Glob to find them
2. Read changed files
3. Load java-25-features.md and spring-boot-4-patterns.md
4. Run Quick Security + Performance checks
5. Report findings with file:line references

Scenario 2: Security Audit

Context: User requests security review

Actions:

1. Load security-checklist.md
2. Glob for controllers, services, security config
3. Check OWASP Top 10 systematically
4. Focus on: SQL injection, authentication, authorization, secrets
5. Provide detailed security report

Scenario 3: Performance Review

Context: User reports slowness or wants optimization

Actions:

1. Load performance-patterns.md
2. Glob for repositories, services
3. Check for: N+1 queries, missing pagination, no caching
4. Review connection pool and async configuration
5. Suggest specific optimizations with expected impact

Scenario 4: Architecture Review

Context: User wants to validate architecture or refactor

Actions:

1. Load architecture-patterns.md
2. Understand current pattern (Layered? DDD? Hexagonal?)
3. Check for violations (layer bypassing, broken boundaries)
4. Validate Spring Modulith modules if used
5. Suggest architectural improvements aligned with their pattern

Scenario 5: Migration Review

Context: Codebase recently migrated to Spring Boot 4 / Java 25

Actions:

1. Load java-25-features.md and spring-boot-4-patterns.md
2. Glob for all Java files
3. Identify old patterns still in use
4. Prioritize quick wins (records for DTOs, switch expressions)
5. Provide migration roadmap for larger changes

When NOT to Use This Skill

• Kotlin code (this skill is Java-specific)
• Non-Spring Boot frameworks (Micronaut, Quarkus)
• Build/config files only (pom.xml, build.gradle) – unless reviewing dependencies
• Frontend code (React, Angular) – unless Spring MVC templates

Example Usage

User: “Can you review OrderService.java for any issues?”

Assistant:

1. Reads OrderService.java
2. Loads java-25-features.md, spring-boot-4-patterns.md, performance-patterns.md, security-checklist.md
3. Analyzes code against checklists
4. Reports findings:

## Critical Issues

🚨 **Performance - N+1 Query**
- **File**: `OrderService.java:45`
- **Problem**: Lazy loading `order.getItems()` in a loop causes N+1 queries
- **Solution**: Use `@EntityGraph` or fetch join in repository

🚨 **Security - Missing Authorization**
- **File**: `OrderService.java:30`
- **Problem**: `deleteOrder()` has no authorization check
- **Solution**: Add `@PreAuthorize("hasRole('ADMIN') or #order.userId == authentication.principal.id")`

## Warnings

⚠️ **Migration - Old Jackson Import**
- **File**: `OrderService.java:5`
- **Problem**: Using `com.fasterxml..jackson.*`
- **Recommendation**: Update to `tools.jackson.*`

## Suggestions

💡 **Modern Java - Use Record for DTO**
- **File**: `OrderService.java:78-120`
- **Suggestion**: Replace `OrderDTO` class with record to eliminate 42 lines of boilerplate

User: “Review the entire order module for architecture issues”

Assistant:

1. Uses Glob to find all files in order/ module
2. Loads architecture-patterns.md
3. Checks for layer violations, DDD patterns, module boundaries
4. Reports architectural findings with recommendations