create-nerq-agent

Scaffold a trust-verified AI agent project in seconds.

npx create-nerq-agent my-agent

Generates a complete agent project with your choice of framework, Nerq trust verification, GitHub Actions CI, and machine discovery built in.

What you get

my-agent/
├── README.md              # With Nerq trust badge
├── requirements.txt       # Framework + nerq
├── .github/
│   └── workflows/
│       └── trust-check.yml  # GitHub Action for trust checking
├── .well-known/
│   └── agent.json         # Machine discovery (A2A protocol)
├── llms.txt               # AI-readable description
├── src/
│   ├── agent.py           # Framework boilerplate
│   └── tools.py           # Dynamic tool discovery via nerq.resolve()
├── nerq.config.json       # Trust verification config
└── LICENSE

Frameworks

• LangChain — Full agent with tools and memory
• CrewAI — Multi-agent crew setup
• AutoGen — Conversational agent framework
• LlamaIndex — RAG and data agent
• Custom — Minimal OpenAI setup

Dynamic tool discovery

The generated tools.py uses nerq.resolve() to find the best tools for any task at runtime:

import nerq

# Find the best tool for a task
tool = nerq.resolve("search github repos")
# Returns: github-mcp-server (Trust: 83, Grade: A)

# Check trust before using a dependency
result = nerq.preflight("langchain")
# Returns: trust_score, grade, recommendation, CVEs

Every project created with create-nerq-agent makes Nerq API calls from day 1.

Options

npx create-nerq-agent my-agent --framework langchain --skip-prompts
npx create-nerq-agent my-agent --no-badge --no-ci

| Flag | Description | |------|-------------| | --framework <name> | langchain, crewai, autogen, llamaindex, custom | | --skip-prompts | Use defaults, no interactive prompts | | --no-badge | Skip Nerq trust badge in README | | --no-ci | Skip GitHub Action | | --no-trust | Skip trust verification setup |

License

MIT