This repository contains security research, technical analysis, and proof-of-concept demonstrations for multiple LiteLLM security findings identified during source code review and testing.
A logic flaw in LiteLLM's API key validation allows authentication using a leaked SHA-256 key hash instead of the original API key.
Impact
- Authentication bypass
- Account takeover
- Administrative access if an admin key hash is exposed
- Unauthorized LLM usage
- Financial impact through API abuse
Location:
auth-bypass-pass-the-hash/
The custom guardrail HTTP primitives allow unrestricted outbound requests to arbitrary URLs, including internal network resources and cloud metadata services.
Impact
- Internal network scanning
- Access to private services
- Cloud credential theft
- Infrastructure compromise
Location:
ssrf-custom-guardrails/
The custom code validator fails to normalize Unicode before applying blacklist-based security checks.
Python automatically normalizes identifiers during execution, allowing forbidden attributes to bypass validation.
Impact
- Sandbox escape
- Access to restricted Python internals
- Secret exposure
- Potential arbitrary code execution
Location:
unicode-sandbox-escape/
Regaan
- GitHub: https://github.com/regaan
- Website: https://rothackers.com
- ORCID: https://orcid.org/0009-0006-3683-7824
This repository is provided for educational, defensive, and research purposes only.
All testing should be performed only on systems you own or are explicitly authorized to assess.