Multi-Engine Scanner

A unified project security scanner that runs Anchore Grype, Aqua Trivy, Astral Ruff and SQLFluff concurrently, deduplicates their findings, and renders them as branded, reviewable reports.

Distributed on PyPI as multi-engine-scanner; the import package is project_scanner and the console commands are multiscan and multi-engine-scanner.

pip install multi-engine-scanner
multiscan /path/to/your/project

Features

• Multi-Engine Scanning: Runs every available engine concurrently for maximum detection coverage.
• Smart Deduplication: Normalizes CVEs and package identifiers, cross-verifying findings to highlight dual-verified vulnerabilities versus tool-unique findings.
• Multi-Format Reporting:
  • 📊 Interactive HTML Reports built for triage: every row states the severity, the finding in plain language, the affected component and the remediation step. Live search, severity and engine filters, sortable and paginated register, expandable detail, CSV export, print/PDF layout.
  • 🗂️ Master Dashboard (index.html) with a hierarchical left rail: executive overview → consolidated register → per-engine reports → transcript and raw exports.
  • 📝 GitHub-Flavored Markdown Report (ideal for pull requests and documentation).
  • 📄 Structured JSON Dataset (ideal for CI/CD integration).
  • 💻 Colorized Terminal Output (ideal for immediate CLI feedback).
• Run Anywhere: Can be executed from any terminal directory against any project path.

Report hierarchy

Every HTML report is laid out as a four-level review hierarchy:

The Severity Weight Index is 100 × Σ(weight × count) ÷ (10 × total), with weights Critical 10, High 6, Medium 3, Low/Unknown 1 — i.e. 100 means every finding is critical.

Presentation lives in project_scanner/reporters/theme.py (palette, stylesheet, shared document furniture); change the brand tokens there to restyle every report at once.

Design tokens

Severity keeps its own semantic palette (red / amber / ochre / indigo) and is always coded twice — a colour bar on the row plus a text badge — so it never depends on colour alone. Layout is responsive: two columns on desktop, one below 1080px, and under 700px the findings table becomes one card per finding rather than a horizontally scrolling grid.

Severity colours stay semantic (red / orange / amber / blue) so criticality is never carried by the brand hue alone. The logo is an original scan-shield mark — the SVG sources are in assets/, and the reports embed it inline so they never fetch a remote image.

Branding the reports

Reports ship with neutral naming. Point the masthead at your own organisation with environment variables — no code changes needed:

export SCANNER_BRAND_NAME="Your Organisation"
export SCANNER_BRAND_UNIT="Information Security Office"
multiscan /path/to/project

Installation

Requires Python 3.9 or newer on macOS, Linux or Windows.

Step 1 — install the scanner

The scanner itself is pure Python standard library, so this command is the same everywhere:

pip install multi-engine-scanner

Installing into a virtual environment is recommended. The only difference between platforms is how you activate it:

python3 -m venv .venv
source .venv/bin/activate
pip install multi-engine-scanner

py -m venv .venv
.venv\Scripts\Activate.ps1
pip install multi-engine-scanner

Step 2 — install the scanning engines

Each engine is a separate tool the scanner shells out to.Install the ones you care about — any engine that is missing is reported as MISSING_BINARY in the report rather than failing the run.

Ruff and SQLFluff come from PyPI on every platform:

pip install "multi-engine-scanner[engines]"

Grype and Trivy are native binaries, so they differ per platform:

macOS

brew install grype trivy

Linux

# Grype - official install script
curl -sSfL https://get.anchore.io/grype | sudo sh -s -- -b /usr/local/bin

# Trivy - official install script
curl -sfL https://raw.githubusercontent.com/aquasecurity/trivy/main/contrib/install.sh | sudo sh -s -- -b /usr/local/bin

Distribution packages work too. Debian / Ubuntu:

sudo apt-get install wget gnupg
wget -qO - https://aquasecurity.github.io/trivy-repo/deb/public.key | gpg --dearmor | sudo tee /usr/share/keyrings/trivy.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/trivy.gpg] https://aquasecurity.github.io/trivy-repo/deb generic main" | sudo tee -a /etc/apt/sources.list.d/trivy.list
sudo apt-get update && sudo apt-get install trivy

Homebrew on Linux also works: brew install grype trivy.

Windows

With Chocolatey in an elevated PowerShell:

choco install grype trivy

Without a package manager, download the release archives, unzip them, and add the folder to your PATH:

• Grype — https://github.com/anchore/grype/releases (grype_*_windows_amd64.zip)
• Trivy — https://github.com/aquasecurity/trivy/releases (trivy_*_windows-64bit.zip)

# add an unzipped folder to PATH for the current session
$env:PATH += ";C:\tools\grype;C:\tools\trivy"

WSL2 is a good alternative on Windows — install inside the Linux distribution using the Linux commands above and scan your project through the /mnt/c/... path.

Note: releases are built and tested on Linux CI. macOS is used daily for development; Windows is supported on a best-effort basis. Use a modern terminal (Windows Terminal or PowerShell 7) so the colourised CLI output renders correctly.

Step 3 — verify

multiscan --help
grype version
trivy --version
ruff --version
sqlfluff --version

Every engine reports its own status in the generated report, so you can also just run a scan and read the engine coverage cards.

From source

git clone https://gitlab.com/chandrabrt/multi-engine-scanner.git
cd multi-engine-scanner
pip install -e ".[dev,engines]"

You can also invoke it as a module without installing the console scripts:

python -m project_scanner /path/to/project

CLI Usage Examples

Scan Current Directory

multiscan

Scan Specific Project Directory

multiscan /path/to/your/project

Specify Output Directory & Formats

multiscan /path/to/target --out-dir ./security-reports --format html,markdown

Filter Minimum Severity

multiscan /path/to/target --min-severity HIGH

Run Single Scanner Only

multiscan /path/to/target --grype-only
# or --trivy-only / --ruff-only / --sqlfluff-only

Auto-fix Python findings

--fix hands the target to Ruff's autofixer before reporting, so the run both repairs what it safely can and tells you what is left.

multiscan --fix

multiscan /path/to/target --fix

This rewrites source files in place. Commit or stash your work first — the scanner prints a warning but will not stop you. Output looks like:

🛠️  Ruff auto-fixed 4 finding(s); 4 remain for manual review.

By default only fixes Ruff considers safe are applied. Add --unsafe-fixes to apply the rest:

multiscan /path/to/target --fix --unsafe-fixes

Unsafe fixes can change behaviour. Removing an unused assignment, for example, deletes the value it held — PASSWORD = "hunter2" disappears along with the F841 warning. Review the diff.

--fix applies to Ruff only; Grype, Trivy and SQLFluff findings are always reported, never modified. Combining it with --grype-only, --trivy-only or --sqlfluff-only is a no-op and warns.

Choose which Ruff rules apply

Rule selection follows Ruff's rule selection rules. --ruff-select replaces the default set (E,F,S,W,C,I,B,ASYNC) and --ruff-ignore removes rules from it. Both also govern what --fix will touch.

# security rules only (Bandit-derived S rules)
multiscan /path/to/target --ruff-only --ruff-select S

# everything except line-length and import-sorting
multiscan /path/to/target --ruff-ignore E501,I001

# fix import ordering and unused imports, nothing else
multiscan /path/to/target --ruff-select I,F401 --fix

Report the installed version

multiscan --version

Output

Reports are written to <target>/vulnerability_reports/ — <target>\vulnerability_reports\ on Windows — and the location can be overridden with --out-dir:

Check what is installed at any time with:

multiscan --version

License

MIT — see LICENSE.