Agent Types

Which agents are available and when to use each. Match the agent’s tool access to the task requirements.

Quick Selection Guide

Need	Agent Type	Why
Search/read code	Explore	Read-only, fast, cheap (haiku)
Architecture design	Plan	Read-only, thoughtful analysis
Run commands/git	Bash	Shell access only
Multi-step implementation	general-purpose	Full tool access
Claude Code questions	claude-code-guide	Docs and web access
Security review	sdlc:security-reviewer	Vulnerability assessment, secure coding
Code quality review	sdlc:quality-enforcer	Formatting, linting, code style
Code simplification	code-simplifier:code-simplifier	Clarity, consistency, minimalism
Architecture analysis	refactor:architect	Design review, optimization planning
Best practices research	adr:adr-researcher	Codebase analysis + web research
Deep code exploration	feature-dev:code-explorer	Trace execution paths, map architecture
Code review	feature-dev:code-reviewer	Bugs, logic errors, conventions
Source code chunk analysis	swarm:rlm-code-analyzer	Code-aware, scope context, analysis focus
Data/CSV chunk analysis	swarm:rlm-data-analyzer	Column-aware, distributions, statistics
JSON chunk analysis	swarm:rlm-json-analyzer	Schema-aware, structural patterns. Also used by JSONL Log Analyzer
General chunk analysis	swarm:rlm-chunk-analyzer	Haiku, fast, structured JSON output
Synthesize chunk findings	swarm:rlm-synthesizer	Sonnet, aggregation and deduplication

Key rule: Read-only agents (Explore, Plan) cannot edit or write files. Never assign them implementation work.

Built-in Agents

Always available without plugins.

Bash

Task({ subagent_type: "Bash", prompt: "Check git status and show recent commits" })

Tools: Bash only
Best for: Git operations, command execution, system tasks

Explore

Task({ subagent_type: "Explore", prompt: "Find all API endpoints", model: "haiku" })

Tools: All read-only (no Edit, Write, Task)
Model: Haiku (fast, cheap)
Best for: Codebase exploration, file searches, code understanding

Plan

Task({ subagent_type: "Plan", prompt: "Design an OAuth2 implementation plan" })

Tools: All read-only
Best for: Architecture planning, implementation strategies

general-purpose

Task({ subagent_type: "general-purpose", prompt: "Implement caching for the user API" })

Tools: All tools (full access)
Best for: Multi-step tasks, research + action combinations

claude-code-guide

Task({ subagent_type: "claude-code-guide", prompt: "How do I configure MCP servers?" })

Tools: Read-only + WebFetch + WebSearch
Best for: Questions about Claude Code, Agent SDK, Anthropic API

statusline-setup

Task({ subagent_type: "statusline-setup", prompt: "Set up a status line showing git branch and node version" })

Tools: Read, Edit only
Model: Sonnet
Best for: Configuring Claude Code status line

Plugin Agents

Available when the corresponding plugins are installed. Plugin agents use the format {plugin}:{agent-name}.

Plugin Agent Constraints: Some agent frontmatter fields are silently ignored when running as plugin agents (for security reasons). Do not rely on these in plugin agent definitions:

hooks — ignored (cannot execute arbitrary shell commands from plugins)

mcpServers — ignored (MCP server injection not permitted from plugins)

permissionMode — ignored (permission escalation from plugins is blocked)

Fields that are supported in plugin agents: name, description, tools, disallowedTools, model, maxTurns, skills, memory, background, isolation.

Review Agents

Agent	Focus
`sdlc:security-reviewer`	Vulnerabilities, supply chain security, secure coding
`sdlc:quality-enforcer`	Formatting, linting, error handling, documentation
`sdlc:compliance-auditor`	Full SDLC compliance audit
`feature-dev:code-reviewer`	Bugs, logic errors, security, code quality, conventions
`code-simplifier:code-simplifier`	Clarity, consistency, minimalism
`refactor:architect`	Architectural analysis, optimization planning
`human-voice:voice-reviewer`	Content voice and tone

Research Agents

Agent	Focus
`adr:adr-researcher`	Codebase analysis + web search for best practices
`feature-dev:code-explorer`	Trace execution paths, map architecture, dependencies
`feature-dev:code-architect`	Design feature architectures based on existing patterns

Refactoring Agents

Agent	Focus
`refactor:architect`	Architecture analysis and optimization planning
`refactor:refactor-code`	Implementation of architectural optimizations
`refactor:refactor-test`	Test coverage analysis and test generation
`refactor:simplifier`	Code simplification and clarity improvements post-refactor
`refactor:security-review`	Security regression review for refactored code

CI/CD Agents

Agent	Focus
`sdlc:ci-architect`	CI pipeline design and GitHub Actions configuration

Test Generation Agents

Agent	Focus
`auto-harness:test-generator`	Comprehensive test definitions for components

RLM Agents

Content-aware chunk analyzers — the Team Lead selects the analyst based on detected content type. In multi-file directory analysis, different analyst types run simultaneously.

Agent	Focus
`swarm:rlm-code-analyzer`	Source code analysis with function/class scope context, analysis focus (Haiku)
`swarm:rlm-data-analyzer`	CSV/TSV analysis with column distributions and statistics (Haiku)
`swarm:rlm-json-analyzer`	JSON/JSONL analysis with schema patterns and field distributions (Haiku). Also used by the JSONL Log Analyzer skill for log-specific schema-aware analysis
`swarm:rlm-chunk-analyzer`	General-purpose: logs, prose, config, markup (Haiku)
`swarm:rlm-synthesizer`	Aggregate chunk findings into coherent reports (Sonnet)

Do NOT override analyst models. Leave the model parameter unset when spawning — agent definitions default to Haiku, which is correct for structured analysis tasks.