Skip to content

Security: ellmos-ai/worksheet-generator

Security

SECURITY.md

Security Policy

Reporting a Vulnerability

If you find a security vulnerability, please report it responsibly:

  1. Do NOT open a public issue
  2. Use GitHub's private vulnerability reporting
  3. Include: description, steps to reproduce, potential impact

How to Report

  1. Go to: https://github.com/ellmos-ai/worksheet-generator/security/advisories/new
  2. Fill out the form (title, description, severity, affected versions)
  3. Submit privately (not visible to public until disclosed)

Security Scope & Principles

  • Local-First & Client Privacy: Worksheet generation, curriculum matching, and pedagogy workflows operate entirely client-side. No telemetry, cloud logging, or automated egress of student or client data is performed.
  • Fail-Closed Secret & API Protection: Credentials and tokens for optional external APIs (such as WHO ICD-11 API credentials via _tools/icf_fetch.py) are strictly confined to gitignored local configuration (config.local.json) or environment variables, and never bundled in generated worksheets or repositories.
  • Safe Rendering & Content Sanitization: Output renderers (Markdown, HTML, DOCX) perform structured validation against schema.py and prevent script execution or insecure remote inclusions.
  • Zero-Copyleft Guarantee: Generated worksheets and teaching materials remain entirely unencumbered by copyleft licenses and are freely usable by schools, clinics, and educators.

Response SLA & Invariants

  • INV-SLA-10 (48-Hour Response SLA): You will receive an initial response and acknowledgment of your report within 48 hours.
  • INV-SLA-09 (5-Day Vulnerability Triage): Full vulnerability verification, severity assessment, and mitigation plan within 5 business days.
  • Coordinated Disclosure: We kindly request allowing a reasonable remediation window before any public disclosure.

There aren't any published security advisories