Presets, Roles & Paradigms

Rigour uses a hierarchical configuration system that combines Project Roles, Coding Paradigms, and Universal Standards to provide expert-level supervision with zero initial configuration.

The Hierarchy of Rigour

When you run an audit, Rigour assembles your quality gates in this order:

1. Universal Config: Base standards (e.g., forbidding TODOs, base complexity).
2. Project Role (preset): Industry-standard thresholds for your project type (API, UI, etc.).
3. Coding Paradigm (paradigm): Language-agnostic structural rules (OOP, Functional).
4. Local Overrides: Your project's specific rigour.yml settings.

---

Universal Standards (UNIVERSAL_CONFIG)

Every Rigour project starts with these "hygiene" gates:

Gate	Default Value	Description
max_file_lines	500	Maximum allowed lines per file
forbid_todos	true	Prevents merging code with TODO markers
forbid_fixme	true	Prevents merging code with FIXME markers
max_files_changed	10	Safety rail for AI agent "explosions"

---

Project Roles (preset)

Roles define the "personality" of your project. They are detected by the existence of specific files or dependencies.

api (Backend Services)

Optimized for high-reliability, maintainable backend logic.

• Detection: express, nestjs, go.mod, requirements.txt, pyproject.toml, main.go.
• Thresholds:
  • max_file_lines: 400
• Required Docs: docs/SPEC.md, docs/ARCH.md, README.md
• Roadmap: Service layer enforcement (Controllers → Services).

ui (Web/React/Next.js)

Optimized for component-based modularity and JSX complexity.

• Detection: react, next, vue, svelte, tailwind.config.js, vite.config.ts.
• Thresholds:
  • max_file_lines: 300
• Required Docs: docs/SPEC.md, docs/ARCH.md, README.md
• Roadmap: Prop-drilling detection (Max depth 5).

infra (IaC/DevOps)

Focuses on safety and preventing accidental infrastructure destruction.

• Detection: Dockerfile, docker-compose.yml, main.tf, k8s/, helm/, ansible/.
• Thresholds:
  • max_file_lines: 300
• Required Docs: docs/RUNBOOK.md, docs/ARCH.md, README.md

data (Data/ML Pipelines)

Optimized for reproducibility and pipeline clarity.

• Detection: ipynb, spark, pandas, dbt_project.yml, data/.
• Thresholds:
  • max_file_lines: 500
• Required Docs: docs/DATA_DICTIONARY.md, docs/PIPELINE.md, README.md
• Roadmap: Stochastic determinism (seed enforcement) and PII leak detection.

---

Coding Paradigms (paradigm)

Paradigms apply syntax-aware AST rules. Rigour scans your source code content to detect the dominant paradigm.

oop (Object-Oriented)

• Patterns: class, interface, extends, constructor, private, public.
• AST Gates:
  • Complexity: 10
  • Max Methods: 10 per class
  • Max Params: 5 per method
  • Inheritance Depth: 3
  • Class Dependencies: 5

functional

• Patterns: export const, reduce(, .pipe(, compose(, curry(, readonly.
• AST Gates:
  • Complexity: 8
  • Max Functions: 15 per file
  • Max Params: 4 per function
  • Max Nesting: 3
  • Function Lines: 40

---

Auto-Discovery

The rigour init command performs a deep scan of your environment:

1. Dependency Scan: Checks package.json, go.mod, etc., for Role markers.
2. Filesystem Scan: Looks for config files (e.g., Dockerfile) for Role markers.
3. Content Heuristics: Samples top source files to detect coding patterns (e.g., heavy use of class vs. const) to assign a Paradigm.

To override discovery:

rigour init --preset api --paradigm oop