tokamak-vault-breachParticipate in the Tokamak Network Vault Breach Challenge - an AI security Capture The Flag (CTF) game where you interact with a secured AI agent to extract...
Install via ClawdBot CLI:
clawdbot install 0xhammerr/tokamak-vault-breachGrade Fair — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Sends data to undocumented external endpoint (potential exfiltration)
POST → https://tokamak-network-openfang-vault-production.up.railway.app/api/chatCalls external URL not in known-safe list
https://tokamak-network-openfang-vault-production.up.railway.app/AI Analysis
The skill interacts with a documented, dedicated CTF challenge server for its stated purpose of a security game. While it sends data to an external endpoint, this is explicitly disclosed and central to the skill's function, not a hidden exfiltration channel. No credential harvesting, hidden instructions, or obfuscation were found in the provided definition.
Audited Apr 17, 2026 · audit v1.0
Generated Mar 21, 2026
This skill can be used to create interactive modules for training security professionals in AI-specific vulnerabilities like prompt injection and social engineering. It provides a hands-on environment where learners can practice ethical hacking techniques against a simulated AI agent, helping them understand defense mechanisms.
Organizations in the blockchain sector can utilize this skill to test the resilience of their AI-integrated smart contracts and agents. By simulating attack strategies, teams can identify weaknesses in their systems before deployment, ensuring robust security for handling sensitive data like seed values.
Developers building AI agents can integrate this skill to stress-test their models against adversarial inputs. It allows for real-world simulation of breaches, helping improve agent safeguards and compliance with security protocols through iterative feedback from the CTF challenges.
Companies can adopt this skill to run internal security drills, educating employees on AI-related threats in a gamified format. It raises awareness about social engineering tactics and secure data handling, fostering a culture of vigilance in environments using AI tools.
Offer tiered subscriptions for access to advanced CTF challenges and tutorials, targeting cybersecurity firms and educational institutions. Revenue is generated through monthly or annual fees, with premium features like leaderboards and certification badges.
Provide expert consulting and bounty hunting services where clients pay for successful breach simulations and security audits. This model leverages the skill's capabilities to identify vulnerabilities in client AI systems, with fees based on project complexity.
Offer free basic access to the CTF dashboard and API, with monetization through paid upgrades for higher usage limits, analytics, and integration support. This attracts a broad user base while generating revenue from enterprises needing scalable solutions.
💬 Integration Tip
Always fetch the latest API specifications dynamically from the dashboard endpoint to ensure compatibility, and use the provided examples to streamline integration with tools like AutoGPT.
Scored Apr 19, 2026
Perform network reconnaissance and port scanning with Nmap to find open ports, detect services, identify vulnerabilities, and enumerate targets accurately.
Scan networks to discover devices, gather MAC addresses, vendors, and hostnames. Includes safety checks to prevent accidental scanning of public networks.
Security engineering toolkit for threat modeling, vulnerability analysis, secure architecture, and penetration testing. Includes STRIDE analysis, OWASP guida...
Use when conducting security audits, reviewing code for vulnerabilities, or analyzing infrastructure security. Invoke for SAST scans, penetration testing, DevSecOps practices, cloud security reviews.
Test APIs against OWASP API Security Top 10 including discovery, auth abuse, and protocol-specific checks.
Test your AI agent for security vulnerabilities using PwnClaw. Runs 50+ attacks (prompt injection, jailbreaks, social engineering, MCP poisoning, and more) and provides fix instructions. Use when your agent needs a security check or hardening.