Skip to content

Security: AI-BuildInfra/proof-graph

Security

SECURITY.md

Security Policy — ProofGraph

ProofGraph is committed to ensuring high security, rigorous SSRF prevention, and safe data retrieval.

1. SSRF Protection Policy

ProofGraph implements strict Server-Side Request Forgery (SSRF) controls:

  • Disallowed Local & Loopback Hosts: localhost, 127.0.0.0/8, ::1.
  • Disallowed Private IP Ranges (RFC 1918):
    • 10.0.0.0/8
    • 172.16.0.0/12
    • 192.168.0.0/16
  • Disallowed Cloud Metadata Endpoints:
    • 169.254.0.0/16 (e.g. 169.254.169.254 AWS/GCP/Azure instance metadata)
    • metadata.google.internal
    • instance-data
  • Redirect Revalidation: Any HTTP 3xx redirect is re-evaluated against the SSRF policy before following.

2. Crawler & Network Guardrails

  • Robots.txt Adherence: Respects robots.txt disallows and crawl-delays.
  • Request Limits: Enforces 8-second request timeouts and 1MB maximum payload limits.
  • Header Identification: Sends explicit user-agent identifying ProofGraph bot as a neutral evidence collector.

3. Reporting Vulnerabilities

To report a security concern, contact security@aibuildinfra.com.

There aren't any published security advisories