kalibur 0.1.5

The AI whitebox pentesting assistant, built by pentesters for engagement workflows

kalibur

The AI whitebox pentesting assistant, built by pentesters for engagement workflows.

Kalibur scans a codebase the way a skilled pentester would: reading every file, tracing data flows from entry points to sinks, and identifying complex exploitable paths across authentication, authorisation, injection, and cryptographic weaknesses. Findings are significantly deeper than what automated scanners surface.

Results land in an editable report.md, structured and ready for manual review. Kalibur does not replace the pentester: it handles the time-consuming groundwork so you can focus on validation, context, and client communication. Triage findings, add notes, mark issues resolved. When you are done, kalibur report proofreads the content, generates an executive summary, and produces a branded PDF in seconds.

The full engagement lifecycle is covered: save and restore snapshots between sessions or across teammates with push and pull, and cleanly close out an engagement with end, which permanently deletes all associated data.

Requirements

• Python 3.9 or later
• A Kalibur API key (kalibur.ai)

Installation

pip install kalibur

Use pip3 or python3 -m pip if pip points to Python 2 on your system:

pip3 install kalibur
# or
python3 -m pip install kalibur

If the kalibur command is not found after installation, run it as python3 -m kalibur.

Authentication

kalibur login

Prompts for your API key, validates it, and saves it to ~/.config/kalibur/key. You can also pass -k <key> to any command or set the KALIBUR_API_KEY environment variable.

Commands

kalibur scan

Run a whitebox security assessment of a codebase.

kalibur scan
kalibur scan -t /path/to/project

Flag	Description
-t PATH	Directory to scan (default: .)
-k KEY	API key

Results are written to a timestamped folder inside a kalibur/ directory in the scanned project:

<target>/kalibur/
  2026-05-03_14-23-45/
    report.md

Add kalibur/ to your .gitignore to keep scan output out of version control.

kalibur report

Generate a branded PDF pentest report from a completed report.md.

kalibur report \
  -r ./kalibur/2026-05-03_14-23-45/report.md \
  -f "Scapin Ltd" \
  -a "Jane Doe" \
  -c "Acme Corp" \
  -l ./logo.png \
  -v "1.0"

Flag	Description
-r FILE	Path to report.md (required)
-f NAME	Assessor firm name (required)
-a NAME	Assessor name (required)
-c NAME	Client name (required)
-l FILE	Logo file, .png or .jpg (required)
-v VER	Report version, e.g. 1.0 (required)
-d	Dark theme (default: light)
-k KEY	API key

The PDF is saved alongside report.md as report-<timestamp>.pdf.

Before generating, review findings in report.md and set each triage status:

# Triage
Status: Open       # finding needs fixing
Status: Accepted   # risk accepted, no fix planned
Status: Resolved   # finding has been fixed

kalibur push

Save a snapshot of your local kalibur/ folder to the cloud.

kalibur push

kalibur pull

Restore a saved snapshot to your local kalibur/ folder.

kalibur pull
kalibur pull -e <project>

Flag	Description
-e PROJECT	Project name (skip interactive picker)

kalibur end

End an engagement and permanently delete all its data: remote scans, snapshots, and the local kalibur/ folder. This cannot be undone.

kalibur end