Agents Reference

Complete reference for all specialized agents available in Everything Claude Code.

What Are Agents?

Agents are specialized subagents that handle delegated tasks with focused expertise. Each agent has:

• Specific tools they can use
• Model selection (Opus for complex reasoning, Sonnet for general work)
• Domain expertise for particular tasks

Trophy Testing Agent

trophy-guide

Trophy Testing specialist that verifies implementations against OpenSpec specs.

Purpose:

• Ingest OpenSpec artifacts (spec.md files with WHEN/THEN scenarios)
• Ingest source code implementation
• Ingest code map (CODEMAP.md) for context
• Use tree-sitter MCP to verify dependency chains
• Generate integration tests that verify WHEN/THEN scenarios
• Run tests to verify implementation correctness

When to use:

• PROACTIVELY after code is written
• To verify implementations match specifications
• Before committing feature code

Trophy Testing Philosophy:

• Implementation FIRST, then verification
• Integration tests are PRIMARY (70%+)
• Minimal mocking - test real behavior
• Spec scenarios become test cases

Planning Agents

planner

Expert planning specialist for complex features and refactoring.

Purpose:

• Analyze requirements and create detailed implementation plans
• Break down complex features into manageable steps
• Identify dependencies and potential risks
• Suggest optimal implementation order
• Consider edge cases and error scenarios

When to use:

• PROACTIVELY for complex feature requests
• Before architectural changes
• For multi-file refactoring

architect

System design specialist for architectural decisions.

Purpose:

• Design system architecture
• Evaluate trade-offs between approaches
• Plan for scalability and maintainability
• Review existing architecture

When to use:

• Major system design decisions
• Evaluating technology choices
• Planning large-scale changes

Code Review Agents

code-reviewer

General code review for quality, security, and maintainability.

Purpose:

• Review code for quality issues
• Check security concerns
• Evaluate maintainability
• Suggest improvements

Review Levels:

• CRITICAL - Must fix before merge
• HIGH - Should fix
• MEDIUM - Consider fixing

When to use:

• PROACTIVELY after writing code
• Before committing changes
• For TypeScript/JavaScript projects

go-reviewer

Go-specific code review specialist.

Focus Areas:

• Idiomatic Go patterns
• Concurrency safety (goroutines, channels, mutexes)
• Error handling patterns
• Performance considerations

When to use:

• After writing Go code
• For Go projects specifically

python-reviewer

Python-specific code review specialist.

Focus Areas:

• PEP 8 compliance
• Type hints and annotations
• Pythonic idioms
• Security best practices

When to use:

• After writing Python code
• For Python projects specifically

security-reviewer

Security vulnerability detection and remediation specialist.

Focus Areas:

• OWASP Top 10 vulnerabilities
• Secret exposure detection
• Input validation
• Authentication/authorization flaws
• Injection vulnerabilities (SQL, XSS, SSRF)

When to use:

• PROACTIVELY after writing auth code
• When handling user input
• Before commits with sensitive data handling

database-reviewer

PostgreSQL database specialist.

Focus Areas:

• Query optimization
• Schema design
• Security
• Performance tuning
• Supabase best practices

When to use:

• Writing SQL queries
• Creating migrations
• Designing schemas
• Troubleshooting database performance

Build & Error Resolution Agents

build-error-resolver

General build error resolution specialist.

Purpose:

• Analyze build error messages
• Identify root causes
• Apply minimal fixes
• Verify build succeeds

When to use:

• When builds fail
• Type errors in TypeScript
• Compilation errors

go-build-resolver

Go build, vet, and compilation error resolution specialist.

Focus Areas:

• Go build errors
• go vet issues
• Linter warnings
• Minimal, surgical fixes

When to use:

• When Go builds fail
• After go vet reports issues
• To fix linter warnings quickly

Testing Agents

e2e-runner

End-to-end testing specialist using Playwright.

Purpose:

• Generate E2E tests for critical user flows
• Manage test journeys
• Quarantine flaky tests
• Upload artifacts (screenshots, videos, traces)

When to use:

• Critical user journey testing
• Full flow verification
• Visual regression testing

Maintenance Agents

refactor-cleaner

Dead code cleanup and consolidation specialist.

Purpose:

• Run analysis tools (knip, depcheck, ts-prune)
• Identify dead code
• Remove unused exports
• Consolidate duplicates

When to use:

• Code maintenance tasks
• Before major refactoring
• Reducing bundle size

doc-updater

Documentation and codemap specialist.

Purpose:

• Update CODEMAP.md files
• Generate documentation
• Keep READMEs current

When to use:

• After significant code changes
• When documentation is outdated

Agent Selection Guide

Using Agents

Agents are automatically invoked by commands or can be called via the Task tool:

# Via commands
/trophy          # Invokes trophy-guide
/go-review       # Invokes go-reviewer
/plan            # Invokes planner

# Directly in conversation
"Use the security-reviewer agent to audit this code"

Agents run with their specified tools and model, returning results to the main conversation.