Skip to main content

Philosophy

The no-bullshit approach to AI code review.

The Problem​

AI coding agents are powerful, but they can also be dangerous. They might:

  • Modify files they shouldn't touch
  • Introduce security vulnerabilities
  • Make changes that break your build
  • Touch too many files at once

The Rigour Solution​

Rigour acts as a safety layer between AI agents and your codebase. It validates every change before it's applied.

Core Principles​

1. Trust, But Verify​

AI agents are helpful. But every change should be validated before it touches production code.

2. Fail Fast​

Catch problems immediately, not after code is merged.

3. Zero Config by Default​

Rigour works out of the box with sensible defaults. Configure only what you need.

4. Transparency​

Every validation result includes clear explanations. No black boxes.

How It Works​

graph TD
    A[AI Agent Task] --> B(Audit: Rigour check)
    B --> C{Pass?}
    C -->|Yes| D[Finish Audit]
    C -->|No| E[Generate Fix Packet]
    E --> F[Agent self-corrects]
    F --> A
  1. Initial Pass: Your agent (CLI or IDE-based) proposes changes.
  2. Audit: Rigour runs internal gates (AST, Safety, Content) and external commands (Tests, Lint).
  3. Diagnostic: If gates fail, Rigour generates a structured Fix Packet (JSON-based engineering feedback).
  4. Self-Correction: The agent consumes the Fix Packet to correct its own violations automatically.

No Bullshit Means​

  • No telemetry - Your code stays private
  • No cloud dependency - Runs entirely local
  • No complex setup - One command to start
  • No false promises - We catch what we can, honestly