Skip to content

Security

Awnexa runs with your AWS credentials, and AI agents such as Claude Code read its output, so we take reports about it seriously. For what Awnexa calls, stores and sends, see the Security overview.

Email security@awnexa.net. Please do not post security issues in public places such as social media or forums.

Please include:

  • The Awnexa version (awnexa version), Node.js version and OS
  • What an attacker can do, and the steps or a proof of concept to reproduce it
  • Whether the issue is already public or known to others

Do not include AWS credentials, account IDs or real configuration data from your environment. Use synthetic data in a proof of concept.

  • Reports are read by a person, not handled by an automated agent.
  • We aim to acknowledge a report within 5 business days and to share an initial assessment within 14 days.
  • We will keep you updated, agree on a disclosure date with you, and credit you in the release notes unless you ask us not to.
  • Fixes are released as a new version and announced in the changelog.

In scope — anything that breaks what Awnexa promises, for example:

  • Awnexa calling an AWS API that is not read-only, or sending requests to an endpoint other than AWS’s own
  • AWS credentials, secrets or masked ([REDACTED]) values being written to the database, Evidence files, logs or output
  • Data being sent anywhere other than Awnexa’s own backend, or usage statistics containing AWS configuration data, account IDs, file paths or host names (see Data sent)
  • Local data files being created with permissions readable by other users (POSIX)
  • Injection through AWS-sourced values that breaks the structure of the JSON, Markdown or Mermaid output
  • The MCP server (awnexa mcp) exposing anything beyond its read-only tools
  • Vulnerabilities in the npm package, its release process or the bundled Claude Code Skills

Out of scope:

  • Misconfigurations in your AWS account that Awnexa reports (that is what Awnexa is for)
  • An AI agent following instructions written in your own AWS resource names or tags. This is a known limitation; please report it only if Awnexa itself makes it worse
  • Vulnerabilities in dependencies that do not affect Awnexa at runtime (report those upstream)
  • Attacks that need write access to your machine or your Awnexa data directory

Until 1.0, only the latest released version receives security fixes.

The contact details are also published in /.well-known/security.txt (RFC 9116).