Skip to content

Security: SlideRuleEarth/sliderule-schema-server

Security

SECURITY.md

Security policy

Reporting a vulnerability

If you believe you've found a security issue in sliderule-schema-server — whether in the served content, the merge pipeline, the Terraform infrastructure, or the packaged Claude skill — please do not open a public GitHub issue.

Instead, email security@mail.slideruleearth.io with:

  • A short description of the issue
  • Steps to reproduce
  • The impact you believe it has
  • (Optional) A suggested fix

We'll acknowledge within 5 business days and keep you updated as we triage, fix, and disclose.

Credible reports that demonstrate exploitable impact will be acknowledged in the release notes unless you prefer otherwise.

Scope

In scope

  • The schema distribution at https://schema.testsliderule.org/ and any production variant.
  • The merge pipeline (schema-endpoints/merge.py and the tree it produces).
  • Terraform in terraform/ — the CloudFront / S3 / ACM / Route 53 wiring.
  • The sliderule-schema Claude skill in skills/sliderule-schema/.

Out of scope

  • Rate-limiting / denial-of-service. CloudFront has standard WAF thresholds; reports of "I can make it slow with 10k requests/s" are expected behaviour.
  • Issues in SlideRule itself — use the sliderule repository instead.
  • Issues in the narrative docs at docs.slideruleearth.io — report to the web-client / docs repos.

What we protect

  • No secrets in the repo. The git history is audited (gitleaks, run against the whole history) and CI includes drift checks for committed artifacts. If you spot something that looks like a leaked credential in the repo, please email rather than post publicly; we'll rotate immediately.
  • Terraform + lock file commit. terraform/.terraform.lock.hcl pins the AWS provider SHA so fresh clones reproduce the same provider build.
  • Static-only serving surface. CloudFront serves JSON from a private S3 bucket; there is no server-side execution, no auth cookie path, no query-string logic.

There aren’t any published security advisories