# Pentest AI

Open-source AI pentester that proves every finding. Machine oracles re-run each exploit; verified bugs ship a proof capsule you can replay yourself.

Repository: https://github.com/0xSteph/pentest-ai
Canonical: https://ross.abutalabs.com/products/pentest-ai
Homepage: https://pentestai.xyz
Language: Python
License: MIT
License Family: permissive
Topics: ai-security, bug-bounty, claude, ctf, cybersecurity, hacking-tools, mcp, model-context-protocol, offensive-security, penetration-testing, pentesting, python, red-team, security, vulnerability-scanner, ai-penetration-testing, ai-pentesting, appsec, devsecops, ethical-hacking
Last push: 2026-08-20T10:32:22+00:00
Link (homepage): https://pentestai.xyz
Link (site_page): https://pentestai.xyz/docs/getting-started

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 98, longevity 10
- inputs: {"age_days": 151, "days_push": 13, "days_rel": 17, "gap_med": 1.0, "n_releases_24m": 23}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1629, forks 314 (observed 2026-08-28T04:05:13.286401+00:00)

## What it is
pentest-ai is an MIT-licensed local CLI and MCP server that turns Claude Code (or any LLM) into an offensive security assistant, pairing 50 specialized subagents with a verification engine that re-runs exploits to prove findings. It runs entirely on your machine with local SQLite storage, no telemetry, and can also be installed as a Claude Code plugin.

## Use cases
- run authorized penetration tests with AI assistance
- verify and reproduce security findings automatically
- plan and manage pentest engagements from the terminal
- integrate pentesting tools into Claude Code via MCP
- research exploits and build detection rules
- generate penetration test reports
- practice for OSCP or bug bounty hunting

## When to choose
- you want AI-assisted pentesting that runs locally with no cloud dependency or telemetry
- you need machine-verified findings rather than unverified scanner flags
- you already use Claude Code, Cursor, or another MCP-compatible client
- you want specialized subagents covering recon, web, AD, cloud, mobile, and post-exploitation

## When to avoid
- you need a fully autonomous scanner with no LLM subscription or API key
- you lack explicit authorization to test your targets - this is for authorized engagements only
- you need a commercial compliance-grade scanning platform with vendor support
- you want a GUI-only workflow without terminal or MCP interaction

## Facets
- artifact type: cli-tool
- maturity: active
- function: penetration-testing, security, agent-framework, mcp, cli, vulnerability-scanning, osint, reverse-engineering
- domain: security, penetration-testing, developer-tools
- platform: cli, python, self-hosted
- tags: claude-code, subagents, red-team, bug-bounty, ctf, mitre-attack, exploit-verification, mcp-server, offensive-security, report-generation, command-line, ai-agents, linux, macos, docker

## Member repositories
- 0xSteph/pentest-ai (main) score 80
- 0xSteph/pentest-ai-agents (plugin) score 72

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:13.286401+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:49:58.281214+00:00, confidence not recorded.
  - readme: https://github.com/0xSteph/pentest-ai (fetched 2026-08-28T04:05:13.286401+00:00, sha c35fcd30d21d)
  - homepage: https://pentestai.xyz (fetched 2026-08-29T10:29:50.184692+00:00, sha 14915fa9030b)
  - site_page: https://pentestai.xyz/docs/getting-started (fetched 2026-08-29T10:29:50.187570+00:00, sha 4ecc21915914)
- Data as of 2026-08-30T08:39:29.467469+00:00.