Quick Start Guide

Quick Start Guide

Get Praetor running on your codebase in under 2 minutes.

Prerequisites

• Access to a long-context LLM (Claude, GPT-5, Gemini, or similar)
• Your source code — as a public GitHub URL, uploaded files, or attached context

Option A: Install Into Your Agentic LLM (Recommended)

npx praetor-audit-kit --install

This auto-detects and installs Praetor into all supported agentic tools on your machine (Claude Code, OpenCode, Cursor, and more). Then simply tell your LLM:

audit my codebase for production readiness

To install into a specific tool only:

npx praetor-audit-kit --install "OpenCode"

Option B: Run Standalone via npm

npx praetor-audit-kit

Launches the interactive scope selector. Pick a scope, review the agents and areas covered, then point it at your codebase.

Other CLI commands:

npx praetor-audit-kit --scope biz          # show Business scope details
npx praetor-audit-kit --scope 2-4          # show range of scopes
npx praetor-audit-kit --check              # pre-flight checks
npx praetor-audit-kit --update             # update to latest from npm
npx praetor-audit-kit --uninstall          # remove from all tools

Option C: GitHub Release

1. Go to Releases
2. Download praetor-prompt-kit-v2.9.2.zip
3. Extract and run npx praetor-audit-kit --install from the extracted directory

After the Discovery Report

Once Phases 0–2 complete (~30 seconds), Praetor stops with a MUST CONFIRM block. Reply:

continue with:
  Q1 = "<your tech stack>"
  Q2 = "<compliance requirements>"
override:
  RUN_PRIORITIES = [P0, P1]
then continue

For your first run, narrow scope with RUN_MODULES = [single-module] and RUN_PRIORITIES = [P0] to see how Praetor works before running the full audit.

Next Steps

• Why Praetor? — Understand the architecture
• The 18 Expert Agents — Meet the specialists
• Priority Rubrics & Overrides — Customize your run
• The 7 Execution Phases — Full pipeline walkthrough