Skip to content

Security: HelloblueAI/bleujs-reasoning-lab

SECURITY.md

Security Policy

Supported versions

Security fixes are applied to the default branch (main) and deployed to production at agi.bleujs.org. Older tags and archived code under src/archive/ are not actively supported.

Version Supported
main (latest) Yes
Older releases Best effort

Reporting a vulnerability

Please do not open a public GitHub issue for security vulnerabilities.

Report security issues through one of these channels:

  1. GitHub Security Advisories (preferred): Create a private advisory

  2. Email: info@helloblue.ai with subject BleuJS Reasoning Lab security

Include:

  • Description of the issue and potential impact
  • Steps to reproduce (proof of concept if available)
  • Affected endpoints, files, or deployment configuration

We aim to acknowledge reports within 72 hours and will coordinate disclosure with you before any public fix.

Secrets and API keys

  • Never commit .dev.vars, wrangler.production.toml, API keys, Cloudflare tokens, or account/KV/zone IDs. Deployment configs belong in the gitignored wrangler.production.toml (template: wrangler.example.toml).
  • Production secrets are set via wrangler secret put (see README.md and .dev.vars.example).
  • pnpm install enables a pre-commit hook (.githooks/pre-commit) that refuses private files and runs gitleaks when installed. CI scans every pull request with the same rules (.gitleaks.toml).
  • If you accidentally commit a secret, rotate it immediately and notify maintainers.

Scope

In scope:

  • The production worker (src/worker/index.ts) and its public API
  • Authentication, authorization, and secret handling in the lab stack
  • Dependency vulnerabilities affecting production deploys

Out of scope:

  • Social engineering against maintainers
  • Denial-of-service against third-party services (e.g. BleuJS API, Cloudflare)
  • Issues in archived code under src/archive/ unless they affect production

Thank you for helping keep BleuJS Reasoning Lab safe for everyone.

There aren't any published security advisories