Skip to content

Releases: blackbuble/pentestai

v0.2.0-beta.3

v0.2.0-beta.3 Pre-release
Pre-release

Choose a tag to compare

@blackbuble blackbuble released this 31 Oct 02:39

PentestAI v0.2.0-beta.3

Highlights

Precision Recall F1 FP

New Features

  • Stored XSS detection — two-pass: inject markers → crawl → verify reflections
  • --cookies — authenticated scanning through entire pipeline
  • --rate-limit — WAF-safe scanning with configurable per-host delay
  • SPA form rendering — Playwright-based form discovery for Angular/React/Vue
  • Ground-truth benchmark — 100% precision, 73.7% recall across 10 targets

Key Fixes

  • CORS/CSRF false positives eliminated
  • SQLi scanner timeout: 973s → 60s cap
  • Interleaved per-param scanning prevents phase starvation
  • ThreadPoolExecutor timeout on all scanners

Install

git clone https://github.com/blackbuble/pentestai.git
cd pentestai && bash setup.sh
source .venv/bin/activate
pentestai autopilot http://target.com --report report.html

Full changelog: CHANGELOG.md

🤖 Generated with Claude Code

v0.2.0-beta.2 — Production-Ready Beta

Pre-release

Choose a tag to compare

@blackbuble blackbuble released this 29 Oct 14:53

PentestAI v0.2.0-beta.2

AI-powered penetration testing framework. 53 tools, 9,000+ CVE templates, zero-config AutoPilot.

Highlights

  • AutoPilot — URL in, full vulnerability report out. Zero config.
  • 4 Scan Presets — quick (~20s), owasp10, api, full
  • 53 Security Tools — SQLi, XSS, SSTI, SSRF, XXE, CORS, CSRF, path traversal, command injection, and more
  • CVE Detection — 30 built-in templates (Log4Shell, Spring4Shell, ProxyLogon, Heartbleed, etc.) + Nuclei integration (9,000+)
  • Professional Reports — HTML with executive summary, risk matrix, remediation plan (20 vuln types)
  • Interactive Guided Wizard — pentestai guided for new users
  • HTTPS Intercepting Proxy — SSL MITM, cert gen, history, replay, HAR export
  • Request Repeater — craft, compare, fuzz params, multi-step sequences
  • Passive Scanner — 22 rules, background traffic analysis
  • WebSocket & gRPC Testing — fuzzing, injection, auth testing
  • Exploit Scripting Framework — write custom exploits in Python
  • TUI — Terminal UI with dashboard, findings, proxy, tools tabs
  • Self-Heal Agent — auto-diagnoses and patches runtime errors

What's New Since v0.2.0-beta

  • Scan presets (--preset quick/owasp10/api/full)
  • Report auto-save (--report report.html)
  • Interactive guided wizard (pentestai guided)
  • Auto-update CVE templates (24h check on scan start)
  • Security audit — 5 vulnerabilities patched (command injection, path traversal, arbitrary exec)
  • Per-request timeout — scans no longer hang on slow targets
  • CVE false positive elimination — all 30 templates fingerprint-tuned
  • Multi-target validated — tested on 5 diverse targets (IIS, Tomcat, Node.js), 0 false positives

Install

# pip
pip install -e .
pentestai guided

# Docker (recommended for full stack)
cp .env.example .env
docker compose --profile msf up -d

# One-liner scan
pentestai autopilot http://target.com --preset owasp10 --report report.html

Stats

Metric Value
Tools 53 (40 available without external binaries)
CVE Templates 30 built-in + 9,000+ via Nuclei
Tests 116 passing
Scan Presets 4 (quick, owasp10, api, full)
Report Formats HTML, JSON, text

Known Limitations

  • SPA/JS-rendered apps need Playwright integration for full form discovery
  • Tested on 5 target types; real-world mileage will improve accuracy
  • Some scanners may trigger WAF on production targets — use --preset quick first

For authorized penetration testing only. Always obtain written permission before scanning any target.

v0.2.0-beta — AI-Powered Penetration Testing Framework

Choose a tag to compare

@blackbuble blackbuble released this 28 Oct 14:10

What's New in v0.2.0-beta

🤖 AI-Powered Features

  • AutoPilot — Give it a URL, get a full vulnerability report. Zero config.
  • SelfHeal Agent — Captures tool errors, diagnoses via LLM, generates patches, tests, and auto-applies fixes.
  • ReAct Agent — Autonomous penetration testing loop with reasoning.

🔍 CVE Detection

🛡️ 43 Security Scanners

  • SQL Injection — Error-based, boolean-based, time-based blind, auth bypass
  • XSS — Reflected + browser execution proof via Playwright
  • SSTI — 7 template engines (Jinja2, Twig, Freemarker, Mako, Smarty, Velocity, ERB)
  • Command Injection — Echo marker + blind timing detection
  • Path Traversal — LFI/RFI, PHP wrappers, null byte, double encoding
  • SSRF — Cloud metadata, internal endpoints
  • XXE — File read, SSRF via entity
  • CORS — Reflected origin, null origin, wildcard detection
  • CSRF — Form token analysis
  • Open Redirect — Host-validated, bypass payloads
  • Header Security — HSTS, CSP, X-Frame-Options, cookie flags, info disclosure
  • IDOR, JWT, OAuth2, GraphQL, Deserialization, File Upload
  • Spider/Crawler — Form discovery, parameter extraction, tech fingerprinting
  • Parameter Fuzzer — 6 wordlists, anomaly detection

🖥️ CLI

pentestai autopilot http://target.com          # Full auto scan
pentestai scan sqli http://target.com/login    # Single scanner
pentestai cve http://target.com --tags apache  # CVE check
pentestai fuzz http://target.com/search?q=FUZZ # Fuzzing
pentestai tools                                # List all 43 tools

📊 Stats

  • 43 registered tools
  • 30 CVE detection templates
  • 23,688 lines of code
  • 1,447 tests passing
  • Rate limiting across all scanners (polite scanning)
  • ASP.NET ViewState auto-extraction

⚠️ Beta Notice

This is a beta release for testing and feedback. Not recommended for production engagements without manual verification of findings.


Full Changelog: https://github.com/blackbuble/pentestai/commits/v0.2.0-beta