👻 Bullshitdetector

Blazing-fast static analysis tool for detecting magic numbers, hardcoded values, and code smells in Rust (and other languages).

🚀 Features

• ⚡ Instant Scanning - Pure regex-based detection, no ML overhead
• 🎯 Magic Number Detection - Finds hardcoded thresholds, timeouts, and constants
• 🔬 Golden Ratio Math - Confidence scoring using φ (phi) = 1.618
• 🌀 Möbius Topology - Advanced code structure analysis
• 📊 PAD Valence Model - Emotional analysis of code patterns
• 🔧 Zero Config - Works out of the box

📦 Installation

As a CLI tool:

cargo install bullshitdetector

As a library:

[dependencies]
bullshitdetector = "0.1"

🎯 Quick Start

Scan for magic numbers:

bullshitdetector scan-magic ./src

Scan for code smells:

bullshitdetector scan ./src --output report.json

As a library:

use bullshitdetector::{DetectConfig, scan_code};

let code = r#"
    if confidence > 0.85 {  // ⚠️ Magic number!
        do_something();
    }
"#;

let config = DetectConfig::default();
let alerts = scan_code(code, &config)?;

for alert in alerts {
    println!("Found {} at line {}", alert.issue_type, alert.location.0);
}

🔍 What It Detects

📊 Example Output

🔴 CRITICAL: Hardcoded Threshold
  File: src/pipeline/stages.rs:231
  Code: if knot > 0.4
  Suggestion: Move to PipelineConfig::knot_threshold
  
🟡 MEDIUM: Magic Number Assignment
  File: src/detection.rs:145
  Code: let base_top_p = 0.35;
  Suggestion: Extract to constant or config

🛠️ Configuration

Create a .bullshitdetector.toml:

[detect]
confidence_threshold = 0.618  # Golden ratio inverse
max_snippet_length = 500
enable_regex_fallback = true

[scan]
exclude_patterns = ["**/test/**", "**/tests/**"]
include_extensions = ["rs", "py", "js"]

🎓 How It Works

1. Regex Pattern Matching - Lightning-fast detection of common patterns
2. AST Analysis - (Optional) Tree-sitter based deep inspection
3. Golden Ratio Scoring - Confidence scoring using φ = 1.618034...
4. Möbius Transforms - Non-orientable topology for code structure
5. PAD Valence - Pleasure-Arousal-Dominance emotional model

🔧 Advanced Usage

Custom Patterns:

use bullshitdetector::{DetectConfig, BullshitType};

let mut config = DetectConfig::default();
config.enable_tree_sitter = false;  // Regex only for speed
config.confidence_threshold = 0.7;   // Adjust sensitivity

let alerts = scan_code(code, &config)?;

Shell Script Integration:

#!/bin/bash
# Pre-commit hook
./bullshitdetector scan-magic src | grep -q "CRITICAL" && exit 1

📚 API Documentation

Full documentation available at docs.rs/bullshitdetector

🤝 Contributing

Contributions welcome! The detector is designed to be extended with new patterns.

Adding Custom Patterns:

1. Add to BullshitType enum in src/lib.rs
2. Implement detection in src/detect.rs
3. Add test cases
4. Submit PR!

📜 License

MIT License - Copyright (c) 2025 Jason Van Pham (ruffian-l on GitHub) @ The Niodoo Collaborative

🙏 Credits

• Golden Ratio Math - Based on φ = (1 + √5) / 2
• Möbius Topology - Non-orientable surface theory
• PAD Model - Pleasure-Arousal-Dominance emotional valence

🔗 Links

• Repository
• Issues
• Changelog

Who you gonna call? 👻 BULLSHITBUSTERS!