An open-source, stateless, and high-performance Model Context Protocol (MCP) framework designed to validate, sanitize, and authorize LLM-generated API payloads.
This project is developed and maintained by SEOSiri-Official (Official website: seosiri.com). To fund ongoing safety research or scale advanced compliance modules, consider supporting the team via the SEOSiri Sponsors Page.
- Universal Core: OWASP injection scanning (SQLi, Command Injection) and credential redaction.
- Fintech (PCI-DSS): Real-time credit card (PAN) detection using the Luhn algorithm.
- Healthcare (HIPAA/FHIR): Automatic scrubbing of PII/PHI (Social Security Numbers, Dates of Birth).
- SEO & Digital Marketing (GDPR/AEO): IP logging redaction and validation of programmatic SEO metadata boundaries (canonical URL structures, title, and description lengths) to protect search index health.
- Biorobotics Safety Interlock: Scans mechanical G-code commands generated by actuators against physical deck limits and safe velocity envelopes, issuing dynamic remediation recommendations.
- Install Package in Editable Mode:
pip install -e .