Skip to content

QUICKSTART.md

Latest commit

 

History

History
92 lines (61 loc) · 2.21 KB

QUICKSTART.md

File metadata and controls

92 lines (61 loc) · 2.21 KB

GRITS Agent Scanner -- Quickstart

Get from zero to a security score in under 5 minutes.

Step 1: Clone and scan (30 seconds)

git clone https://github.com/X-Scale-AI/grits-agent-scanner.git
cd grits-agent-scanner

# OpenClaw
./grits-agent-scanner

# NVIDIA NemoClaw
./grits-agent-scanner --agent nemoclaw

No dependencies. Python 3 stdlib only.

The scanner reads your actual config files and checks 20 things across 5 security layers. It tells you in plain language what each failure means, what you can fix now, and what needs an expert. Nothing is modified.

Step 2: Preview the fixes (1 minute)

./grits-agent-secure

Shows you exactly what will change, creates a backup location, and prints the rollback command. Nothing is applied until you confirm.

Step 3: Apply fixes (2 minutes)

./grits-agent-secure --apply

The fixer will:

  1. Show what it will change
  2. Create a full backup with a rollback command
  3. Ask for confirmation
  4. Apply only if you say yes

Covers: firewall, tool deny list, sandbox mode, file permissions, dangerous flags.

Step 4: Re-scan to verify

./grits-agent-scanner

Your score should go up. Remaining findings are the ones that need human judgment (network segmentation, secrets architecture, identity setup, audit logging).

Step 5: Manual hardening

For the issues the auto-fixer cannot handle:

Option A: Do it yourself using the guides in apply/openclaw/ or apply/nemoclaw/.

Option B: Talk to an expert. Schedule a free 30-min call: https://xscaleai.com/consult

Step 6: Host hardening (optional, 10 minutes)

# Linux host (CIS Benchmark + DISA STIG)
sudo bash tools/harden.sh

# Docker host
sudo bash tools/harden-docker.sh

Review the configuration section at the top of each script before running.

Output formats

# Terminal (default) -- color output, designed for screenshots
./grits-agent-scanner

# Markdown -- for PRs, wikis, security reports
./grits-agent-scanner --report > security-report.md

# JSON -- for dashboards, CI/CD, automation
./grits-agent-scanner --json > scan-results.json

Scored with GRITS v0.3.0 by X Scale AI https://github.com/X-Scale-AI/grits-agent-scanner | https://xscaleai.com