Attacker obtains code execution in proxy VM mitigated

THREATQW23 · Mitigated: Controls in place bring the risk down to an acceptable level. · Defined in catalog.py:3344

Risk
High impact × Medium likelihood = 9 high
After mitigations
High impact × Low likelihood = 6 medium

How statuses and scores work

An attacker that gains code execution in the sd-proxy VM can intercept and modify traffic between the client running in sd-app and the Securedrop instance's journalist interface. An attacker could delete sources, submissions or download encrypted submissions and replies, or send messages to sources. An attacker may also exfiltrate the journalist interface ATHS token.

Threat actors: Nation State / Law Enforcement

Mitigations

Mitigation Status
qrexec that proxies API requests and responses only to SecureDrop application server MITIGQW19 implemented

Related threats

This threat is part of: Vulnerability introduced in provisioning code THREATQW15

Smaller threats that make up this one:

Where it applies

Applies to 1 component in 1 scenario.

Workstation: reading submissions · sd-proxy

Notes

Existing mitigations: - Submissions, messages and replies go through sd-proxy encrypted with the gpg encryption key. - securedrop.Proxy qubes-rpc policy only allows sd-app, and requests from all other AppVMs are denied. - We wrote a qrexec service that proxies API requests only to the Journalist Interface (MITIGQW19). Further mitigations: sd-proxy uses a whonix-workstation template, and may present extra attack surface.

How this threat is matched

Listed for these components. A boundary covers every component inside it.