This project executes user-specified FPGA toolchains inside Docker/Podman containers and adheres to zero-trust principles:

• SLSA build provenance (L2) for the published release archive (the release zip is the attestation subject).
• Deterministic Builds ensuring verifiable and reproducible binary outputs.
• Non-root Container Execution preventing namespace privilege escalation.
• Static Analysis enforced via GitHub CodeQL and Trivy.

If you discover a security vulnerability, please report it responsibly by emailing info@mtorun0x7cd.com instead of opening a public issue.

You should receive an acknowledgement within 48 hours. Critical issues affecting container isolation or credential exposure will be prioritized for immediate out-of-band patches.

Security-relevant areas in-scope include:

• Container escape - Any bypass of the Docker sandbox boundary via extension mounts or config.
• Credential leakage - Exposure of host environment variables, SSH keys, or socket paths via telemetry or logs.
• Path traversal - Workspace mount injection that accesses files outside the active project directory.
• Zombie Process Exhaustion - Unhandled PIDs overloading host PID limits.

Two settings deliberately relax the isolation boundary. Both default to disabled, and a bypass reached through either is not a vulnerability:

• Allow Native Fallback - Runs the tool directly on the host when no container engine is reachable, bypassing container isolation by design.
• Allow Privileged Containers - Permits --privileged, forfeiting the --cap-drop=ALL and no-new-privileges posture of the default path.

Issues related to host-side Docker daemon misconfiguration (e.g., exposing unauthenticated TCP sockets) or inherent zero-days in the EDA tools themselves (e.g., buffer overflows in yosys) are outside the scope of this project, provided they do not facilitate a container escape.