create-agent

Generate custom agent prompts, command launchers, and persistent memory files.

520|175|Updated Apr 8, 2026
One-click install
npx skills add https://github.com/evolution-foundation/evo-nexus --skill create-agent-evolution-foundation
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: create-agent
Source: https://github.com/evolution-foundation/evo-nexus/tree/main/.claude/skills/create-agent
Command: npx skills add https://github.com/evolution-foundation/evo-nexus --skill create-agent-evolution-foundation

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Create-agent helps you set up a new workspace-specific custom agent (with the right prompt, command, and memory) instead of manually scaffolding the files each time.

Core Features & Use Cases

  • Agent scaffolding: Creates a .claude/agents/custom-{name}.md file with system instructions for domain, personality, model, and dashboard color.
  • Workspace integration: Adds a slash command in .claude/commands/custom-{name}.md that launches the correct subagent type and ties to the agent’s persistent memory.
  • Persistent memory setup: Creates a dedicated .claude/agent-memory/custom-{name}/ folder for durable agent-specific notes.
  • Proper safety/consistency guardrails: Guides you to defer to config/workspace.yaml for language/owner/company and to follow naming and folder conventions.
  • Core-agent overlap awareness: Advises when not to duplicate existing core agents and when to narrow the custom scope.

Quick Start

Tell Claude Code: create a new agent for "customer support" with a friendly tone, use model sonnet, pick a blue hex color, and store outputs under workspace/support/.

Frequently Asked Questions about create-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a custom agent for a recurring workflow in Claude Code?

To create a custom agent, you scaffold the agent prompt, command launcher, and persistent memory by placing files under `.claude/agents/`, `.claude/commands/`, and `.claude/agent-memory/` using a `custom-` naming convention.

What is the best way to set up persistent memory for a multi-agent workspace?

Setting up persistent memory requires creating a dedicated `.claude/agent-memory/custom-{name}/` folder, which stores durable agent-specific notes and ties directly to the agent's slash command launcher.

How does agent scaffolding integrate with workspace configuration?

Agent scaffolding integrates with workspace configuration by deferring to `config/workspace.yaml` to determine language, owner, and company context, ensuring the custom agent follows proper workspace safety and consistency guardrails.

Can I use this to duplicate existing core agents for domain-specific tasks?

You should avoid duplicating existing core agents; instead, the agent creation process advises narrowing the custom scope to prevent overlap and ensure the new agent addresses a specialized domain like finance operations or DevOps monitoring.

Do I need to manually create slash commands when generating a new workspace agent?

You do not need to manually create slash commands, because the scaffolding process automatically adds a `.claude/commands/custom-{name}.md` file that launches the correct subagent type and connects it to persistent memory.

What configurations are required to define a custom agent's behavior and appearance?

Configuring a custom agent requires defining its domain instructions, personality, model type, and dashboard color within the `.claude/agents/custom-{name}.md` file during the scaffolding process.