Skip to content

Security: blondres04/shieldclaw

Security

SECURITY.md

Security Policy

Supported Versions

Currently, only the main branch is receiving security updates. Shield Claw is in V0 Proof of Concept stage.

Reporting a Vulnerability

If you discover a vulnerability in the Shield Claw architecture—such as a sandbox escape, a telemetry injection flaw, or a JWT authentication bypass—do not open a public GitHub issue.

Please report it privately by emailing the repository maintainer. You will receive an acknowledgment within 48 hours. We will coordinate a fix and disclosure timeline with you before making any public announcement.

Known Security Considerations (V0 Architecture)

Shield Claw is actively working to harden its execution environment. Users should be aware of the following accepted architectural risks in the V0 release:

  1. Docker Socket Mounting: The Sandbox Orchestrator mounts the host's Docker socket (/var/run/docker.sock) to spin up the target containers. This means the target containers run as siblings on the host daemon, not as true nested Docker-in-Docker. This is a known privilege escalation surface. Do not run Shield Claw on a host containing sensitive production workloads.
  2. Outbound Network Access: Sandbox containers currently require outbound network access to perform git clone operations on the target repository. While ephemeral, they are not completely air-gapped.
  3. Default Credentials: The Docker Compose stack ships with default JWT credentials (auditor / secure2026). These must be rotated if deployed outside of a local, air-gapped testing environment.

There aren't any published security advisories