Skip to content

Security: tuneay/Redline

SECURITY.md

Security policy

Supported versions

Redline is currently pre-1.0. Security fixes are applied to the latest commit on the main branch.

Reporting a vulnerability

Do not disclose vulnerabilities, leaked credentials, prompt contents, full conversations, or sensitive logs in public issues, discussions, or pull requests.

Use GitHub's private vulnerability reporting flow:

  1. Open the repository's Security tab.
  2. Select Advisories.
  3. Choose Report a vulnerability.
  4. Include reproduction steps, impact, affected files or routes, and any suggested mitigation.

Use redacted examples. Never include real API keys, personal data, or unnecessary conversation content.

Relevant reports include:

  • server-side secret exposure;
  • client-bundle leakage of protected prompts or credentials;
  • prompt-injection paths that override system rules;
  • stored or reflected script injection;
  • local-session corruption with a security impact;
  • denial-of-service or request-limit bypasses;
  • dependency vulnerabilities with demonstrated impact on Redline.

Maintainers will investigate valid reports privately and coordinate disclosure after a fix is available. Response times are best-effort because this is a community-maintained project.

Security model

  • Groq credentials remain server-only.
  • User messages are treated as untrusted prompt data.
  • API input, model output, evidence, and local storage are schema-validated.
  • Production code does not log full conversations.
  • There is no database, account system, analytics collector, or web-search integration.

Credential exposure

If a secret is exposed, revoke or rotate it immediately. Removing it from the latest commit is not sufficient because Git history and external caches may retain it.

There aren't any published security advisories