Skip to content

Agent Catalog

36 specialized agents organized by domain. Each agent is an expert in its area — call the right one for the job.

Interactive Catalog

Browse all agents and skills with search, filters, and details in the interactive catalog.

Core Agents

The foundation. Building, reviewing, debugging, and migrating code.

AgentDescriptionExample
@starterCreate new projects from scratch"Use @starter to create a Next.js SaaS with Stripe"
@builderBuild modules, components, services"Use @builder to create the products module with CRUD"
@reviewerUnified 3-in-1 code review (quality, security, architecture)"Use @reviewer to review src/modules/auth/"
@doctor4-phase systematic debugging"Use @doctor to investigate the checkout timeout"
@migratorModernize legacy code safely"Use @migrator to convert class components to hooks"

Workflow Agents

Orchestrating how you work — planning, executing, testing, pairing.

AgentDescriptionExample
@plannerAdaptive planning that scales with complexity"Use @planner to plan the notification system"
@executorExecute plans with checkpoints and rollback"Use @executor to implement the plan"
@tddTest-Driven Development with Red-Green-Refactor"Use @tdd to build the payment validator"
@debuggerSystematic debugging with hypothesis testing"Use @debugger to trace the memory leak"
@pairAI pair programming with real-time feedback"Use @pair to work on the search feature"
@analystTransform requirements into technical specs"Use @analyst to spec the multi-tenant system"
@orchestratorCoordinate multiple agents for complex tasks"Use @orchestrator to build the auth module"
@scoutAnalyze project structure and recommend agents"Use @scout to analyze this project"
@memoryPersist decisions and context across sessions"Use @memory to save the architecture decisions"

Engineering Agents

Deep domain expertise — from API design to cloud architecture.

AgentDescriptionExample
@apiREST/GraphQL API design with OpenAPI specs"Use @api to design the orders endpoint"
@perfPerformance profiling and optimization"Use @perf to optimize the dashboard load time"
@i18nInternationalization and localization"Use @i18n to add Spanish support"
@docsGenerate documentation from code"Use @docs to document the auth module"
@refactorSafe code refactoring with tests"Use @refactor to extract the validation logic"
@depsDependency management and upgrades"Use @deps to audit and update dependencies"
@explorerDeep codebase exploration and analysis"Use @explorer to map the data flow in checkout"
@financePayment systems, billing, subscriptions"Use @finance to implement Stripe subscriptions"
@cloudCloud architecture, IaC, serverless"Use @cloud to design the AWS infrastructure"
@securityAuth, OWASP, encryption, audit"Use @security to audit the authentication flow"
@designerDesign systems, accessibility, UI patterns"Use @designer to create the component library"
@dataDatabase design, migrations, optimization"Use @data to design the multi-tenant schema"
@devopsDocker, Kubernetes, CI/CD pipelines"Use @devops to set up the GitHub Actions pipeline"
@testerTest strategies and coverage plans"Use @tester to create the testing strategy"
@legalGDPR, LGPD, compliance implementation"Use @legal to implement LGPD consent management"
@architectFull system architecture design and migration"Use @architect to migrate to hexagonal architecture"
@rippleImpact analysis of code changes"Use @ripple to analyze the impact of removing UserService"

Business Agents

Strategy, growth, and user-facing operations.

AgentDescriptionExample
@marketingCopy, SEO, growth experiments, social strategy"Use @marketing to write landing page copy"
@productProduct roadmaps, specs, PRDs, feature prioritization"Use @product to create the Q1 roadmap"
@supportSupport docs, FAQs, runbooks, escalation paths"Use @support to create the API troubleshooting guide"

Automation Agents

Automate repetitive ops — error triage, production monitoring, fix workflows.

AgentDescriptionExample
@sentry-triagePull Sentry errors, cross-reference PRs, prioritize by severity, auto-create fix PRs"Use @sentry-triage to check errors from the last 24h"
@autopilotIterative autonomous builds with PRD and progress tracking"Use @autopilot to build the feature from docs/PRD-auth.md"

Framework-Specific Agents

Each framework pack includes optimized versions of core agents:

PackAgents Included
Vue 3@builder, @reviewer, @doctor, @migrator
React@builder, @reviewer, @doctor, @migrator
Next.js@builder, @reviewer, @doctor, @migrator
SvelteKit@builder, @reviewer, @doctor, @migrator
Angular@builder, @reviewer, @doctor, @migrator
Astro@builder, @reviewer, @doctor, @migrator
Nuxt@builder, @reviewer, @doctor, @migrator

Agent Modes

Every agent has two modes:

ModeModelUse Case
FullSonnet/OpusMaximum quality, detailed output
LiteHaikuLower cost, faster responses

Choose during installation: npx specialist-agent init

Agent Frontmatter

Agents support advanced frontmatter fields for Claude Code integration:

FieldValuesPurpose
namestringAgent identifier (required)
descriptionstringCSO-optimized "Use when..." (required)
toolscomma-separatedAllowed tools
modelsonnet, opus, haiku, inheritModel override
effortlow, medium, high, maxReasoning effort level
skillslistPre-loaded skills for the agent
memoryuser, project, localCross-session persistence
maxTurnsnumberLimit agent turns
isolationworktreeRun in isolated git worktree
colorhexUI display color

TIP

effort, memory, skills, maxTurns, and isolation are officially supported by Claude Code 2.1.75+. Plugin agents cannot use hooks, mcpServers, or permissionMode for security reasons.