Presets

Arxo includes pre-configured presets for common analysis scenarios. Presets select appropriate metrics and set reasonable defaults.

Available Presets

Quick

Use Case: Fast feedback during development

Metrics:

SCC (cycle detection)
L1 Overview (high-level stats)

Characteristics:

⚡ Very fast (1-5 seconds)
🎯 Essential metrics only
💡 Quick health check

Usage:

arxo analyze --quick
# or
arxo analyze --preset quick

Example Output:

✓ 183 files analyzed
✓ No circular dependencies
✓ System size: 247 modules

CI

Use Case: Continuous integration pipelines

Metrics:

SCC
Propagation Cost
Smells
Hierarchy

Characteristics:

⚡ Fast (10-30 seconds)
🚫 Fail-fast on violations
📊 Essential quality gates

Usage:

arxo analyze --preset ci --fail-fast

Recommended CI Config:

preset: ci

policy:
  invariants:
    - metric: scc.max_cycle_size
      op: "=="
      value: 0

    - metric: propagation_cost.system.ratio
      op: "<="
      value: 0.20

    - metric: smells.cycles.severe_count
      op: "=="
      value: 0

run_options:
  fail_fast: true
  quiet: true

Risk

Use Case: Identify technical debt and architectural risks

Metrics:

SCC
Propagation Cost
Smells
Centrality
MSR (code churn, hotspots)
Change Coupling

Characteristics:

🔍 Comprehensive risk analysis
📈 Evolution metrics
🎯 Identifies refactoring candidates
⏱️ Medium speed (30-60 seconds)

Usage:

arxo analyze --preset risk

What It Finds:

Circular dependencies
God components (high coupling)
Change hotspots (frequently modified)
Coupled changes (files that change together)
Fragile modules (high propagation cost)

Coupling

Use Case: Analyze dependencies and coupling

Metrics:

propagation_cost
msr (hotspot and co-change analysis)
coupling_analysis
decoupling_level
clustered_cost_cochange
semantic_duplicates

Characteristics:

🔗 Dependency focus
📊 Coupling metrics
🎯 Refactoring guidance
⏱️ Medium speed (30-60 seconds)

Usage:

arxo analyze --preset coupling

What It Measures:

Direct coupling (imports)
Transitive coupling (propagation)
Co-change coupling from git history
Duplication pressure from semantically similar modules

Security

Use Case: Security-focused analysis

Metrics:

SCC
Security Vulnerabilities
Sensitive Data Flow
Effect Violations

Characteristics:

🔒 Security focus
🕵️ Data flow analysis
⚠️ Risk detection
⏱️ Medium speed (30-90 seconds)

Usage:

arxo analyze --preset security

What It Detects:

Sensitive data leaks
Unsafe data flows
Missing input validation
Side effect violations

Runtime

Use Case: Runtime and operational analysis

Metrics:

Runtime Centrality
Traffic Hotspots
Critical Path
Runtime Drift
Sensitive Data Flow (prioritizes PII violations on hot paths when telemetry is present)
Test Coverage (prioritizes undertested modules on hot paths when telemetry is present)
Chaos Readiness
Cloud Cost

Characteristics:

📊 Operational focus
🔍 Production insights
💰 Cost analysis
⏱️ Requires telemetry data

Usage:

arxo analyze --preset runtime --telemetry-path ./traces.json

Or configure telemetry in your config file (see Configuration):

data:
  telemetry:
    source_path: ./traces.json
    format: otel_json   # or zipkin_json, jaeger_json

Requirements:

Trace data in one of: OTLP JSON, Zipkin JSON, or Jaeger JSON
Set data.telemetry.format to match your exporter
For span-to-code mapping: include code.filepath in span attributes/tags

See the Telemetry guide for collection and format details.

AI

Use Case: AI/LLM application analysis

Metrics:

LLM Integration
RAG Architecture
Agent Architecture
ML Architecture
Fine-tuning Architecture

Characteristics:

🤖 AI-specific patterns
💰 Cost tracking
🔒 PII detection
📊 Observability gaps
⏱️ Medium speed (30-60 seconds)

Usage:

arxo analyze --preset ai

What It Detects:

LLM calls without tracing
Missing cost tracking
PII leakage risks
Hardcoded prompts
Missing evaluation harness
Rate limit absence

Full

Use Case: Comprehensive analysis (all metrics)

Metrics: All 50+ available metrics

Characteristics:

📊 Complete analysis
🔬 Maximum insight
⏱️ Slow (2-5 minutes)
💾 Large output

Usage:

arxo analyze --preset full

When to Use:

Initial architecture audit
Quarterly reviews
Before major refactoring
Comprehensive documentation

Combining Presets with Custom Metrics

You can extend presets:

# Start with CI preset, add custom metric
arxo analyze --preset ci --metric llm_integration

Or in config:

preset: ci

metrics:
  # Preset metrics + custom
  - id: llm_integration
  - id: rag_architecture

Creating Custom Presets

While you cannot register new presets, you can create reusable configs:

.arxo.custom.yaml:

# Custom "frontend" preset
metrics:
  - id: scc
  - id: propagation_cost
  - id: modularity
  - id: centrality
  - id: smells

data:
  import_graph:
    group_by: folder
    group_depth: 2
    exclude:
      - "**/test/**"
      - "**/*.test.tsx"

policy:
  invariants:
    - metric: scc.max_cycle_size
      op: "=="
      value: 0
    - metric: propagation_cost.system.ratio
      op: "<="
      value: 0.15

run_options:
  fail_fast: true

Use:

arxo analyze --config .arxo.custom.yaml

Preset Comparison

Preset	Speed	Metrics	Use Case
quick	⚡⚡⚡ Very Fast (1-5s)	2	Development loop
ci	⚡⚡ Fast (10-30s)	4	CI/CD gates
risk	⚡ Medium (30-60s)	6	Tech debt audit
coupling	⚡ Medium (30-60s)	6	Dependency analysis
security	⚡ Medium (30-90s)	4	Security review
runtime	⚡ Medium (varies)	6	Operational analysis
ai	⚡ Medium (30-60s)	5	AI/LLM projects
full	🐌 Slow (2-5min)	50+	Comprehensive audit

Preset Selection Guide

For Development

Pre-commit: --quick
IDE integration: --quick
Manual checks: --preset ci

For CI/CD

Pull Requests: --preset ci --only-changed
Main branch: --preset risk
Nightly: --preset full

For Reviews

Sprint review: --preset risk
Architecture review: --preset coupling
Security audit: --preset security
Quarterly audit: --preset full

For Specialized Projects

AI/LLM apps: --preset ai
Microservices: --preset coupling
Legacy refactor: --preset risk
Production monitoring: --preset runtime

Environment-Specific Presets

Create different configs per environment:

Development (.arxo.dev.yaml):

preset: quick
policy:
  invariants: []  # No enforcement

Staging (.arxo.staging.yaml):

preset: ci
policy:
  invariants:
    - metric: scc.max_cycle_size
      op: "<="
      value: 3  # Allow small cycles

Production (.arxo.prod.yaml):

preset: risk
policy:
  invariants:
    - metric: scc.max_cycle_size
      op: "=="
      value: 0  # Strict

AI-Assisted Preset Selection

Use the MCP server with your AI assistant to discover and run presets conversationally:

You: "What analysis presets are available?"
AI: [Calls list_presets] "8 presets available: quick, ci, risk..."

You: "Run a security analysis"
AI: [Uses analyze_architecture with preset="security"]
    "Found 2 sensitive data flow violations..."

The MCP server supports all presets and can help you choose the right one. See MCP Workflows for examples.

Next Steps

CLI Reference - Command options
Configuration Guide - Full config
MCP Workflows - Preset selection with AI
Policy Examples - Policy patterns
Metrics Overview - Available metrics