DiffCatcher

🎯 Key Features

• Repository Discovery: Recursively scan directories for Git repos with configurable filters
• State Tracking: Capture pre/post-pull state with commit hashes, messages, and dirty detection
• Diff Generation: Automatic N vs N-1 and historical diff creation with file manifests
• Element Extraction: Parse diffs to identify functions, structs, classes, imports, and more across 10+ languages
• Code Snippets: Extract full before/after code with boundary detection and context windows
• Security Tagging: 18 built-in security patterns (crypto, auth, secrets, SQL injection, XSS, etc.)
• Multi-Format Reports: JSON, Markdown, text, and SARIF outputs with cross-repo security overview
• Branch-Diff Mode: Diff any two refs (branches, tags, commits) in a single repo — ideal for PR reviews
• Performance: Parallel processing with progress bars, LRU caching, and incremental mode

Why not just use bash?

A one-liner like ls | while read line; do git -C "$line" diff HEAD~1 HEAD || true; done only shows raw diffs. DiffCatcher adds recursive discovery, code element extraction, security pattern detection, SARIF output for CI/CD, parallel processing, and cross-repo security aggregation. See full comparison below.

📋 Table of Contents

• Installation
• Quick Start
• Usage
  • Basic Scanning
  • Pull Modes
  • Extraction Options
  • Security Tagging
  • Configuration File
  • Plugin System
  • Branch-Diff Mode (PR Review)
  • SARIF Output
  • Advanced Features
• Report Structure
• Configuration
• Architecture
• Testing
• Documentation
• Contributing

🚀 Installation

From Source

git clone https://github.com/Teycir/DiffCatcher.git
cd DiffCatcher
cargo build --release
./target/release/diffcatcher --help

Requirements

• Rust 1.70+
• Git 2.0+

⚡ Quick Start

# Scan all repos in a directory (fetch-only, no modifications)
diffcatcher ~/projects

# Pull updates and generate security report
diffcatcher ~/projects --pull -o ./report

# Diff two branches in a single repo (PR review mode)
diffcatcher ./my-repo --diff main..feature/auth -o ./pr-report

# Generate SARIF output for GitHub Code Scanning
diffcatcher ~/projects --summary-format sarif,json -o ./report

# Dry run to see what would be scanned
diffcatcher ~/projects --dry-run

# Fast scan with 8 parallel workers
diffcatcher ~/projects -j 8 --quiet

📖 Usage

Basic Scanning

# Scan with default settings (fetch-only)
diffcatcher <ROOT_DIR>

# Custom output directory
diffcatcher ~/projects -o ./my-report

# Include nested repos and follow symlinks
diffcatcher ~/projects --nested --follow-symlinks

# Skip hidden directories
diffcatcher ~/projects --skip-hidden

Pull Modes

# Fetch only (default - no working tree changes)
diffcatcher ~/projects

# Actually pull changes
diffcatcher ~/projects --pull

# Force pull with stash/pop for dirty repos
diffcatcher ~/projects --pull --force-pull

# Use rebase strategy
diffcatcher ~/projects --pull --pull-strategy rebase

# Skip fetch/pull entirely (historical diffs only)
diffcatcher ~/projects --no-pull

Extraction Options

# Skip element extraction (raw diffs only)
diffcatcher ~/projects --no-summary-extraction

# Extract elements but skip code snippets
diffcatcher ~/projects --no-snippets

# Adjust snippet context and limits
diffcatcher ~/projects --snippet-context 10 --max-snippet-lines 300

# Limit elements per diff
diffcatcher ~/projects --max-elements 1000

Security Tagging

# Skip security tagging
diffcatcher ~/projects --no-security-tags

# Include test files in security analysis
diffcatcher ~/projects --include-test-security

# Use custom security patterns
diffcatcher ~/projects --security-tags-file ./custom-patterns.json

Configuration File

DiffCatcher can auto-load project-local configuration from:

• <ROOT_DIR>/.diffcatcher.toml (default)
• a custom file via --config <FILE>
• disabled with --no-config

Example:

output = "reports-local"
no_pull = true
history_depth = 2
summary_formats = ["json", "txt"]
no_security_tags = false

[plugins]
security_pattern_files = ["plugins/security-extra.json"]
extractor_files = ["plugins/extractors.json"]

CLI flags still override config values when explicitly set.

Plugin System

DiffCatcher supports two plugin types:

• Security pattern plugins via --security-plugin-file <FILE> (repeatable)
• Extractor plugins via --extractor-plugin-file <FILE> (repeatable)

Security plugin format matches --security-tags-file JSON (version, mode, tags).

Extractor plugin format:

{
  "version": 1,
  "extractors": [
    {
      "name": "policy-rule",
      "kind": "Config",
      "regex": "^policy\\s+([A-Za-z_][A-Za-z0-9_]*)"
    }
  ]
}

Branch-Diff Mode (PR Review)

# Diff two branches in a single repo
diffcatcher ./my-repo --diff main..feature/auth

# Diff specific commits
diffcatcher ./my-repo --diff abc123..def456

# Diff with SARIF output for CI integration
diffcatcher ./my-repo --diff origin/main..HEAD --summary-format sarif -o ./pr-report

The --diff BASE..HEAD flag skips repository discovery and fetch/pull — it directly diffs two refs (branches, tags, or commit SHAs) and runs the full extraction + security tagging pipeline on the result.

SARIF Output

# Generate SARIF alongside other formats
diffcatcher ~/projects --summary-format sarif,json,md

# SARIF-only for CI/CD upload
diffcatcher ~/projects --summary-format sarif -o ./report

When sarif is included in --summary-format, a results.sarif file is written to the report root. This file follows the SARIF 2.1.0 standard and integrates with GitHub Code Scanning, VS Code SARIF Viewer, Azure DevOps, and other SARIF-compatible tools.

Advanced Features

# Incremental mode (skip unchanged repos)
diffcatcher ~/projects --incremental -o ./report

# Filter by branch pattern
diffcatcher ~/projects --branch-filter "main"

# Adjust history depth
diffcatcher ~/projects --history-depth 5

# JSON output for CI/CD
diffcatcher ~/projects --quiet --json > result.json

# Verbose output with discovered paths
diffcatcher ~/projects --verbose

📁 Report Structure