Skip to content

Security: cognitum-one/meta-llm-docs

Security

SECURITY.md

Security Policy

Scope

This repository hosts the public, high-level documentation and landing page for Cognitum MetaHarness-as-a-Service. It is a static site — plain HTML and CSS, with no JavaScript, no analytics, no trackers, no external CDNs, no cookies, and no application code. The product's API, source code, configuration, and credentials are private and auth-gated and are not part of this repository.

Reporting a vulnerability

If you believe you have found a vulnerability, an exposed secret, or any over-exposure in this public site, please report it privately:

  • Email: security@cognitum.one
  • Please include the URL/path, a description, and (if relevant) reproduction steps.
  • Please do not open a public issue for suspected secret exposure.

We aim to acknowledge reports promptly and will coordinate disclosure responsibly.

What this site deliberately does NOT contain

  • No API keys, tokens, HMAC/signing secrets, or secret-manager secret names.
  • No internal service-account identities, project identifiers, or private hostnames beyond the already-public, auth-gated API endpoint name.
  • No private API integration details, request/response payloads, or internal URLs.

The site is intentionally a high-level overview. The private code and private API stay private.

Hardening applied to the published site

  • Strict Content-Security-Policy meta tag (default-src 'self', script-src 'none', no inline or remote scripts; styles served from a self-hosted stylesheet only).
  • referrer: no-referrer and X-Content-Type-Options: nosniff meta directives.
  • HTTPS-only (enforced by GitHub Pages) with upgrade-insecure-requests.
  • No cookies, no client-side fetches, no third-party origins of any kind.
  • A machine-readable /.well-known/security.txt (RFC 9116).

There aren't any published security advisories