@prontiq/ariscan

Measure and improve how ready your codebase is for AI coding agents.

ARI (Agent Readiness Index) is a composite score (0-100) derived from 8 research-calibrated pillars. Run a single command to get an actionable readiness report.

Install

npx @prontiq/ariscan .

Or install globally:

npm install -g @prontiq/ariscan
ariscan .

Usage

The examples below use ariscan, which is available after npm install -g @prontiq/ariscan. Replace ariscan with npx @prontiq/ariscan if you haven't installed globally.

# Scan current directory
ariscan .
# — or, without global install —
npx @prontiq/ariscan .

# Scan a specific path
ariscan /path/to/repo

# JSON output (for CI pipelines)
ariscan . --json

# Fail CI if score is below threshold
ariscan . --threshold 60

# Markdown report (with "Quick Start: Top 3 Actions")
ariscan . --format markdown

# SARIF output for GitHub Code Scanning
ariscan . --format sarif

# Generate SVG badge
ariscan . --badge ari-badge.svg

# Export JSON Schema for output validation
ariscan --json-schema > ariscan.schema.json

Exit codes: 0 = pass, 1 = below threshold, 2 = runtime error.

What It Scores

| Pillar | Weight | What It Measures | |---|---|---| | Agent Context Quality | 15% | AGENTS.md quality, information additionality | | Feedback Loop Speed | 15% | Local test/lint/typecheck speed, CI duration | | Test Isolation | 18% | Cloud credential independence, DI patterns | | Dev Environment | 10% | Devcontainer, bootstrap scripts | | Doc Machine-Readability | 10% | OpenAPI schemas, error taxonomy | | Build Determinism & Type Safety | 15% | Strict mode, lockfiles, reproducible builds | | Code Navigability | 12% | Module boundaries, import graph, complexity | | Security & Governance | 5% (gate) | Branch protection, CODEOWNERS, secrets scanning |

Maturity Levels

| Level | Score | Meaning | |---|---|---| | L1 Hostile | 0-25 | Agents face significant friction — missing guardrails | | L2 Fragile | 26-45 | Simple edits feasible with close supervision | | L3 Capable | 46-65 | Routine tasks with moderate supervision | | L4 Productive | 66-80 | Multi-file features with light supervision | | L5 Autonomous | 81-100 | Complex cross-service tasks, self-verified |

Security acts as a gate: scoring below 40% on Security & Governance caps the overall level at L2.

Programmatic API

For programmatic use, import scan from @prontiq/ariscan-engine directly:

import { scan } from "@prontiq/ariscan-engine";

const result = await scan(".");
console.log(result.score);         // 0-100
console.log(result.level);         // "L1" .. "L5"

Related Packages

• @prontiq/ariscan-schema — Type definitions and Zod schemas
• @prontiq/ariscan-engine — Scoring engine for programmatic use

License

Elastic License 2.0 (ELv2) — free to use, modify, and redistribute. Cannot be offered as a managed service.