If you discover a security vulnerability in this project, please do not open a public issue. Instead, report it privately:

• Email: abhinaykrupa@gmail.com — subject line: [SECURITY] vertical-procurement-toolkit
• GitHub: use private security advisories

Please include:

• A description of the issue and where it lives in the code
• Steps to reproduce
• The impact you believe it has
• Any suggested fix you've identified

I aim to acknowledge reports within 3 business days and provide a remediation timeline within 10 business days.

This toolkit is a reference architecture intended to be embedded in larger systems. Within this scope:

In scope — please report:

• Code execution via crafted CSV input (e.g. through pandas, YAML, or the matcher)
• Path traversal or arbitrary-file-read via the CLI / adapters
• Prompt injection in the real LLM judge (vpt/llm_judge.py) that exfiltrates data or breaks structured output
• Dependency vulnerabilities directly exploitable through the documented entry points

Out of scope — won't be treated as vulnerabilities:

• The mock Stripe/HubSpot data in app/sync/ (it's deliberately fabricated for demo)
• Lack of authentication in the Streamlit demo (it's a POC, see SECURITY_REVIEW.md)
• Mocked LLM judge behavior — this is a development-time stub, not production
• Recommendations to "use environment variables for keys" — already documented

If you embed this toolkit in a production system, the burden of additional hardening is on you. Start with:

• Read SECURITY_REVIEW.md for a full POC-vs-production security gap analysis
• Read PRODUCTION_ARCHITECTURE.md for the recommended production swap-ins (pgvector, real LLM, auth, audit log)
• Add input size limits before passing user files to adapters
• Run the LLM judge with structured output enforcement (already on for OpenAI provider)
• Treat any catalog data as sensitive — it represents negotiated commercial terms