# nuguard

Perform automated behavior validation of your AI Applications and Agents. The validation tests are dynamically generated for red-team tests and happy-path behavior validation.

## Quick Start

```bash
# Connect this server (installs CLI if needed)
npx -y smithery mcp add NuGuardAI/nuguard

# Browse available tools
npx -y smithery tool list NuGuardAI/nuguard

# Get full schema for a tool
npx -y smithery tool get NuGuardAI/nuguard nuguard_init

# Call a tool
npx -y smithery tool call NuGuardAI/nuguard nuguard_init '{}'
```

## Tools (7)

- `nuguard_init` — Initialize nuguard.yaml with auto-detected defaults (SBOM path, policy file, canary seed). Also creates companion start…
- `nuguard_sbom_generate` — Generate an AI Bill of Materials (AI-SBOM) from Python or TypeScript source code, or by cloning a git repository. Detec…
- `nuguard_analyze` — Run static risk analysis on an AI-SBOM. Evaluates 18 NGA structural rules plus MITRE ATLAS mapping, OSV / Grype CVE sca…
- `nuguard_scan` — Run the full unified security scan pipeline: SBOM generation → static analysis → policy check → red-team (optional). Wr…
- `nuguard_behavior` — Run intent-aware behavioral testing against a live AI application endpoint. Static mode checks SBOM–policy alignment; d…
- `nuguard_redteam` — Run adversarial red-team testing against a live AI application. Generates exploit scenarios from the AI-SBOM and execut…
- `nuguard_policy_check` — Cross-check a Cognitive Policy document against an AI-SBOM and run compliance framework assessments (OWASP LLM Top 10, …

```bash
# Get full input/output schema for a tool
npx -y smithery tool get NuGuardAI/nuguard <tool-name>
```