AgentCommunity★ 2k

rshah515 165+ Subagents

Updated Jul 7, 2026

A collection of over 165 Claude Code subagents organized into categories: core, languages (13), frameworks (13), infrastructure (13), data/AI, databases (7), quality (11), mobile, web3, marketing (30), and more. Agents auto-discover on startup from ~/.claude/agents/ or the project .claude/agents/ directory — no manual activation required.

View on GitHub

What it does

architect
Designs system architecture, API contracts, and database schemas with rationale for each decision.
prd-writer
Generates product requirements documents and user stories; recommended starting point for complex projects.
project-manager
Orchestrates multi-agent workflows, breaks down tasks, and manages timelines across specialist agents.
tech-lead
Reviews architecture decisions, sets code standards, and provides senior technical guidance.
incident-commander
Coordinates crisis response and emergency procedures during production incidents.
terraform-expert
Plans and applies Terraform IaC changes across AWS, GCP, and Azure with state management and module design.
security-auditor
Reviews code against OWASP standards and produces a prioritized vulnerability report.

Files (1)

OVERVIEW.mdprimary · markdown · 2.1 KB

## What it does

A collection of 165+ Claude Code subagents organized into 20+ categories covering the complete software development lifecycle. Categories include core development (architect, debugger, code-reviewer, refactorer), 13 language specialists (Python, Go, TypeScript, Rust, Java, Kotlin, Swift, and more), 13 framework experts (React, Next.js, Vue, Django, Rails, FastAPI, NestJS, Svelte, and more), 13 infrastructure agents (Terraform, Kubernetes, Ansible, GitOps, observability), 7 database specialists (PostgreSQL, MongoDB, Redis, Elasticsearch, Neo4j, Cassandra), 11 QA agents (Playwright, Cypress, Jest, k6, chaos engineering), 30 marketing specialists (SEO, GEO, email, content, PR), industry verticals (fintech, healthcare, govtech, edtech), and orchestrators (prd-writer, project-manager, tech-lead, incident-commander).

Agents auto-discover from ~/.claude/agents/ (global) or .claude/agents/ (project-local) on Claude Code startup — no manual activation step needed.

## Key agents

- architect — system design, API architecture, database schemas
- prd-writer — product requirements and user stories; recommended starting point for complex projects
- project-manager — multi-agent orchestration and task breakdown
- tech-lead — architecture decisions and code standards
- incident-commander — crisis response and emergency coordination
- terraform-expert — IaC, multi-cloud deployments, state management
- security-auditor — OWASP, penetration testing, compliance
- owasp-top10-expert — web application security review

## Install

Clone the repo to your Claude agents directory:

    cd ~/.claude/agents
    git clone https://github.com/rshah515/claude-code-subagents comprehensive-agents

Agents activate on the next Claude Code startup. Use explicit invocation ("use the architect to design...") or let Claude auto-route based on context.

## When to use

Use prd-writer first for any complex, multi-component project; then project-manager to coordinate the implementation agents. For single-domain work, invoke a specific expert directly ("have the nextjs-expert refactor this component").