Package Detector (PD) - Universal Security CLI

Package Detector acts as a security gatekeeper for developers, wrapping npm and pip installation commands to analyze packages for malicious behavior before they are installed locally.

Features

• Multi-Registry Support: Automatically handles NPM and PyPI ecosystems.
• Security Analysis: Uses AST parsing and RAG-based threat modeling to detect risky patterns.
• Interactive Prompts: Blocks high-risk packages and asks for user confirmation.
• Premium CLI UX: Uses rich for beautiful reports and questionary for smooth interactions.

Installation

The recommended way to install Package Detector is using the provided installation script, which sets up an isolated virtual environment.

# Clone the repository
git clone https://github.com/teamoptimadev/package-detection-cli.git
cd package-detection-cli

# --- FOR MAC / LINUX ---
bash scripts/install.sh

# --- FOR WINDOWS (PowerShell) ---
powershell.exe -ExecutionPolicy Bypass -File .\scripts\install.ps1

Global One-Line Install (For Other Users)

Users can install Package Detector without cloning the repo using these commands:

macOS / Linux (Bash):

curl -sSL https://raw.githubusercontent.com/teamoptimadev/package-detection-cli/main/scripts/install.sh | bash

Windows (PowerShell):

iwr -useb https://raw.githubusercontent.com/teamoptimadev/package-detection-cli/main/scripts/install.ps1 | iex

Post-installation: If ~/.local/bin is not in your PATH, add it to your profile (~/.zshrc or ~/.bashrc):

export PATH="$PATH:$HOME/.local/bin"

Usage

Basic Installation (Safe)

Intercept any installation command:

pd install react
# or
pd i numpy

Explicit Registry

Force a specific registry if heuristic detection fails:

pd install --npm some-package
pd install --pip some-python-package

Full Proxy Commands

Use pd as a prefix for standard installers:

pd npm install lodash
pd pip install flask

Scanning (No Install)

Analyze a package without triggering the installer:

pd scan --registry npm cross-env

How it Works

1. Command Intercept: pd captures the package name and intent.
2. Download & Extract: Downloads the package metadata and source code to a temporary sandbox.
3. Behavior Extraction: Performs AST analysis to extract sensitive function calls and system interactions.
4. RAG Match: Compares extracted behaviors against a vector database of known malicious patterns.
5. AI Reasoning: A reasoning module summarizes the risk and provides a verdict (SAFE, SUSPICIOUS, MALICIOUS).
6. Gatekeeping: If the verdict is risky, it prompts the user. If verified safe, it executes the system's actual npm/pip binary.

---

Built by Package Detector Team. Stay Safe! 🛡️