Disclaimer
Effective Date: March 1, 2026
Basilisk is an offensive security tool. Unauthorized use against AI systems you do not own or have explicit written permission to test may violate federal and international cybercrime laws.
1. Purpose & Scope
Basilisk ("the Software") is an open-source AI/LLM red teaming framework developed by Regaan under the Rot Hackers research initiative. It is designed exclusively for authorized security testing, vulnerability research, and academic study of artificial intelligence systems.
The Software includes 29 attack modules, a genetic prompt evolution engine (SPE-NL), and multi-provider LLM integration capabilities. These features are intended to help security professionals identify and remediate vulnerabilities in AI systems before malicious actors exploit them.
2. Authorized Use Only
You may use Basilisk only in the following authorized environments:
- AI/LLM systems owned and operated by you
- Systems for which you hold explicit, documented, written authorization from the system owner
- Dedicated testing and staging environments isolated from production
- Academic and educational research environments
- Authorized penetration testing engagements under a signed scope agreement
3. Prohibited Activities
The following uses of Basilisk are strictly prohibited:
- Testing AI systems without explicit written authorization from the system owner
- Extracting, exfiltrating, or disclosing proprietary system prompts, training data, or model weights
- Deploying evolved attack payloads against production AI systems without authorization
- Using the Software to bypass safety guardrails on AI systems serving the public
- Distributing, selling, or weaponizing payloads generated by Basilisk
- Using the Software in violation of any applicable law, regulation, or contractual obligation
4. Tool-Specific Warnings
Prompt Injection Modules
Send adversarial instructions to override LLM guardrails. May cause the target to execute unintended actions or disclose system-level information.
SPE-NL Evolution Engine
Breeds and mutates attack payloads across generations. Evolved payloads may bypass defenses that resist static attacks. Use with extreme caution.
System Prompt Extraction
Designed to extract hidden system prompts. Successful extraction may reveal trade secrets or proprietary instructions.
Multi-Turn Escalation
Simulates social engineering across turns. May cause the AI to gradually relax its safety constraints over an extended conversation.
5. Limitation of Liability
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS, CONTRIBUTORS, OR ROT HACKERS BE LIABLE FOR ANY CLAIM, DAMAGES, OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT, OR OTHERWISE, ARISING FROM, OUT OF, OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
You are solely and exclusively responsible for ensuring your use of Basilisk complies with all applicable laws, regulations, and organizational policies. The developers assume no responsibility for misuse, data loss, service disruption, or legal consequences arising from the use of this Software.
6. Acknowledgement
By downloading, installing, or using Basilisk, you acknowledge that you have read and understood this Disclaimer in its entirety. You agree that you will use the Software only in authorized environments and accept full responsibility for your actions.
Report Security Issues
If you discover a vulnerability in Basilisk itself, please report it responsibly via support@rothackers.com.