PraisonAI MCPSecurity authentication bypass for Basic and OAuth methods

Executive brief

The PraisonAI package contains a security component used to protect AI-related server tools and resources. A flaw in this component allows users to bypass security checks when using 'Basic' or 'OAuth' authentication methods by providing any non-empty credential. This could allow unauthorized individuals to access protected data, list internal tools, or trigger automated workflows without a valid password or token.

Technical details

A protection mechanism failure exists in the `MCPSecurity` helper within the `praisonai` npm package. The `evaluatePolicy()` function in `src/praisonai-ts/src/mcp/security.ts` contains logic that only invokes the `validate` callback for 'api-key' and 'bearer' authentication methods. For 'basic' and 'oauth' methods, the code extracts the token but fails to call the validator, defaulting to an 'allowed' state for any non-empty credential. An unauthenticated remote attacker can exploit this by sending a request with an arbitrary 'Basic' or 'OAuth' Authorization header to bypass intended access controls on MCP servers. The issue is addressed in version 1.7.2.

Affected products

• MervinPraison praisonai >= 1.5.1, <= 1.7.1

Timeline

• 2026-06-17: disclosed
• 2026-06-18: advisory: GHSA-4qq2-2j2x-x62c published
• 1.7.2: patched

References

• https://api.github.com/users/rexpository
• https://github.com/rexpository
• https://api.github.com/users/rexpository/gists%7B/gist_id%7D
• https://api.github.com/users/rexpository/repos
• https://avatars.githubusercontent.com/u/30176934?v=4
• https://api.github.com/users/rexpository/events%7B/privacy%7D