create-agent — Claude Code skill: what it does, how to install
Scaffolds autonomous subagents for Claude Code with valid YAML frontmatter, isolated contexts, and triggering examples.
npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a claude-codeWhat create-agent does
This skill generates configuration files for autonomous Claude Code subagents stored in agents/*.md. Unlike slash commands defined in commands/*.md, agents run as separate subprocesses spawned via the Task tool. The parent conversation invokes them automatically when incoming tasks match the trigger rules outlined in the description.
It enforces Anthropic frontmatter conventions. Agent identifiers require 3 to 50 characters using only lowercase letters, digits, and hyphens. The description must begin with "Use this agent when..." and provide <example> blocks with scenario context and expected replies. Skipping concrete examples causes erratic auto-triggering during active coding sessions.
Frontmatter fields also control UI badge colors, model inheritance (inherit, sonnet, opus, or haiku), and tool access. Restricting privileges to read-only utilities such as Read or Grep protects your codebase from unintended modifications during background reasoning.
What it helps with
- Scaffold subagent configuration files for Claude Code
- Define automatic trigger conditions using structured dialogue examples
- Restrict agent permissions to read-only tools for safer execution
- Assign dedicated UI colors and models to background subagents
Example prompts for the agent
- Create a code-reviewer subagent that triggers when I ask to inspect git diffs
- Generate a security-analyzer agent restricted to Read, Grep, and Glob tools
- Draft an agent file for automated test generation using the sonnet model
How to install create-agent
1Skills CLI (Vercel)
npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a claude-code2Manually: put the files in place
~/.claude/skills/create-agent/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/create-agent/
1Skills CLI (Vercel)
npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a codex2Manually: put the files in place
.agents/skills/create-agent/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a cursor2Manually: put the files in place
.cursor/skills/create-agent/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install NeoLabHQ/context-engineering-kit create-agent2Manually: put the files in place
.github/skills/create-agent/SKILL.md1Skills CLI (Vercel)
npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a gemini-cli2Platform CLI
gemini skills install https://github.com/NeoLabHQ/context-engineering-kit --path skills/create-agentManually: put the files in place
.devin/skills/create-agent/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about create-agent
How to install create-agent in Claude Code?
Run npx skills add NeoLabHQ/context-engineering-kit --skill create-agent -a claude-code. The skill will be placed in ~/.claude/skills/create-agent/SKILL.md.
How do subagents differ from slash commands?
Slash commands in commands/*.md execute directly in the main conversation when invoked by the user. Subagents in agents/*.md run in isolated contexts triggered autonomously by the model.
What naming conventions are enforced for agents?
Names must be 3 to 50 characters long, begin and end with an alphanumeric character, and contain only lowercase letters, numbers, or hyphens. Underscores and uppercase letters are not allowed.
