Skip to content

Back to the catalog

GitHub Secure Code Game
github.blogcourse

GitHub Secure Code Game

Season 4 of GitHub's open-source Secure Code Game puts players inside 'ProdBot', a deliberately vulnerable agentic coding assistant modeled on tools like OpenClaw and GitHub Copilot CLI. Across five progressive levels players must get ProdBot to leak a secret by exploiting sandbox escapes, untrusted web content ingestion, malicious MCP server integrations, poisoned org-approved skills/persistent memory, and multi-agent trust chains. It runs free in GitHub Codespaces and references real-world risk categories from the OWASP Top 10 for Agentic Applications 2026 and CVE-2026-25253 ('ClawBleed') RCE in OpenClaw.

TL;DR Free open-source CTF-style game teaching agentic AI attacks (sandbox escape, prompt injection, MCP tool abuse, memory poisoning) via a vulnerable AI agent, runnable in GitHub Codespaces.
Open original ↗

Organisations named here

  • OWASP Top 10 for Agentic Applications 2026

CVEs named here

  • CVE-2026-25253

Technologies named here

  • GitHub Secure Code Game
  • OpenClaw
  • ProdBot
  • GitHub Codespaces
  • GitHub Copilot CLI
  • Model Context Protocol
  • GitHub

Techniques named here

  • Sandbox breakout via natural language commands
  • Untrusted content / prompt injection via web browsing
  • MCP tool abuse
  • Trust exploitation via skills and persistent memory
  • Multi-agent trust chain exploitation / memory poisoning
  • Agent goal hijacking
  • Tool Misuse
  • Identity abuse
  • Memory poisoning

Related in AI Security