Skip to main content
skill CX Codex

pentest-r1-autonomous-penetration-testing — Claude Code skill

pentest-r1-autonomous-penetration-testing guides setup of a two-stage reinforcement learning pipeline for autonomous penetration testing. Ideal for training security models in interactive sandboxes.

no license 271 installs ★ 13 updated 2026-10-09

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add reason-machines/security-skills --skill pentest-r1-autonomous-penetration-testing -a claude-code
Open on GitHub

What pentest-r1-autonomous-penetration-testing does

This skill guides the setup of LLM training for penetration testing. The first phase runs grpo_stage1.py. It trains a base model via GRPO and LoRA using 500 expert walkthrough records.

The second phase runs grpo_multi_turn_stage2.py inside an intercode-ctf Docker sandbox. A reward class scores command attempts, penalizing errors and giving points when flags are captured.

Executing this pipeline requires Python 3.11.11, Docker, and NVIDIA Container Toolkit. Without proper GPU support, policy optimization scripts cannot run.

What it helps with

  • configure two-stage reinforcement learning for security agents
  • run offline GRPO training on expert pentest logs
  • set up interactive intercode-ctf Docker challenge environments
  • implement reward logic for command outputs and flag discovery

Example prompts for the agent

  • Help me build the intercode-ctf Docker container for online training
  • Configure the Stage 1 LoRA settings and run grpo_stage1.py
  • Show how PentestRewardModel scores CTF command responses

How to install pentest-r1-autonomous-penetration-testing

1Skills CLI (Vercel)

npx skills add reason-machines/security-skills --skill pentest-r1-autonomous-penetration-testing -a claude-code

2Manually: put the files in place

~/.claude/skills/pentest-r1-autonomous-penetration-testing/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/pentest-r1-autonomous-penetration-testing/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about pentest-r1-autonomous-penetration-testing

How do I install pentest-r1-autonomous-penetration-testing in Claude Code?

Run npx skills add reason-machines/security-skills --skill pentest-r1-autonomous-penetration-testing -a claude-code.

What does this skill do?

It provides setup steps and configuration patterns for training security models using offline GRPO on walkthrough data and online interactive CTF Docker sandboxes.

What prerequisites are needed before starting the training scripts?

You need Python 3.11.11, Docker, and NVIDIA Container Toolkit. GPU acceleration also requires the CUDA 12.4 runtime.

Similar skills

All skills for Security →

From the same set129

The whole set →
iac-security-scan-skills iac-security-scan-skills is a Claude Code skill for auditing Terraform and CloudFormation, tracing cross-domain attack paths and calculating minimal fixes. CX Codex +1 319installs linux-pentester-command-reference linux-pentester-command-reference — a skill that hands your agent a Linux command reference split by pentest phase: recon, enumeration, exploitation, privesc. CX Codex +1 313installs linux-pentesting-command-reference linux-pentesting-command-reference — a skill that feeds the agent a Linux command cheat sheet for recon, service enumeration, privilege escalation and staying on the box. Handy in labs and CTFs where a one-liner beats theory. CX Codex +1 290installs microsoft-security-skills-plugin This skill gives AI agents Microsoft Security guidance for Defender, Sentinel, Entra, Purview, and Intune. Use it when planning policies or security workflows. CX Codex +1 276installs solana-security-standard solana-security-standard is a Claude Code skill that detects 37 vulnerability classes in Solana and Anchor programs during development and CI. CX Codex +1 268installs jellyfin-security-plugin jellyfin-security-plugin walks your coding agent through adding two-factor login, passkeys, OIDC single sign-on and IP bans to a Jellyfin 10.11+ server. CX Codex +1 265installs deepseek-pentest-ai-burp-extension deepseek-pentest-ai-burp-extension guides configuration and usage of a DeepSeek-driven Burp Suite extension to automate parameter detection and security fuzzing. CX Codex +1 262installs claude-pentest-framework claude-pentest-framework — a Claude skill that runs a structured pentest with 15 executor agents and 63 attack categories, asking the operator for approval before anything is exploited. CX Codex +1 261installs

Project sponsors

Companies that keep this analytics open