/

CVE-2026-35465

Severity:

HIGH RISK

(7.5)

Path Traversal / Insecure File Write in Gzip Extraction

CVE-2026-35465

Analyzed

Vulnerability type

Category:

Path Traversal / Insecure File Write

CWE:

CWE-36, CWE-73

Impact:

Remote Code Execution via Malicious Archive Overwrite

Severity:

HIGH RISK

(7.5)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

SecureDrop Client versions 0.17.4 and below contain a path traversal flaw in their gzip archive extraction logic that fails to properly validate filenames, allowing absolute paths within archives.

If an attacker has already compromised the SecureDrop Server, they can craft malicious archives that overwrite critical files on the Client's sd-app virtual machine, potentially leading to code execution and full compromise of decrypted source submissions.

Affected context

This vulnerability appears in the SecureDrop Client desktop application when it processes gzip archives received from a SecureDrop Server and extracts them within the sd-app virtual machine.

The issue manifests when archive entry names are not sanitized for absolute paths or traversal sequences, allowing files to be written outside the intended extraction directory.

It becomes relevant in environments where the Client implicitly trusts archives produced by the SecureDrop Server, particularly in high-risk newsroom deployments handling sensitive submissions.

Exploit preconditions

  • Attack surface: Archive processing pipeline within the SecureDrop Client sd-app VM.
  • Required access: Prior compromise of the dedicated SecureDrop Server that generates or transmits gzip archives.
  • Trust boundary: The Client treats archives coming from the Server as trusted content without sufficiently constraining resulting file paths.
  • Trigger point: Automatic or user-initiated extraction of a gzip archive on the Client workstation.
  • Why exploitation is possible: The extraction code does not normalize and restrict archive entry paths to a safe base directory, allowing absolute paths that can overwrite critical application or data files, including the SQLite database, and enabling code execution primitives.

Why this happens

The core issue is insufficient validation and normalization of file paths during gzip archive extraction, a classic instance of CWE-36: Absolute Path Traversal combined with CWE-73: External Control of File Name or Path.

The extraction logic accepts filenames supplied by the archive as-is and writes them to the filesystem without enforcing that they remain under a designated extraction root.

Because the archive contents are treated as trusted and not constrained, an attacker controlling the archive can specify absolute paths or crafted paths that escape the intended directory and overwrite security-sensitive files, creating a path for arbitrary code execution or persistent state manipulation.

Potential impact

  • Overwriting of the SecureDrop Client's SQLite database and other critical application files.
  • Execution of attacker-controlled code or scripts within the sd-app virtual machine context.
  • Compromise of confidentiality for decrypted source submissions and related metadata.
  • Integrity violations, including tampering with stored submissions, logs, or application state.
  • Denial of service by corrupting essential configuration or runtime files needed by the Client.

General mitigation guidance

Effective mitigation focuses on strictly constraining file paths derived from archive contents and hardening the archive processing pipeline:

  • Normalize and validate each archive entry path before extraction, resolving symbolic links and rejecting absolute paths or traversal patterns (e.g., components that escape the designated extraction directory).
  • Enforce an explicit, immutable extraction root directory and ensure all writes are checked against this root after path normalization.
  • Implement allowlisting for permitted file locations and file types, denying extraction of unexpected paths or sensitive system locations.
  • Treat all archive contents from remote services, including hardened servers, as untrusted input and apply the same defensive parsing and validation routines.
  • Run archive extraction under a minimally privileged account or sandboxed context so that, even if a flaw occurs, the blast radius is limited.

How teams detect issues like this

Detection can leverage both preventive and retrospective approaches:

  • Static analysis: Configure rules to flag file extraction code that writes paths directly from archives or external inputs without normalization and base-directory checks.
  • Code review: Pay particular attention to archive handling utilities, looking for use of absolute paths, lack of canonicalization, or missing checks that ensure paths remain under a specific root.
  • Runtime monitoring: Monitor the Client environment for unexpected file writes to sensitive paths during archive processing, especially outside designated data directories.
  • Policy enforcement: Adopt internal secure coding guidelines that mandate path normalization and directory confinement for all file operations driven by remote content; integrate these checks into CI pipelines and automated review tools.

Sources

SourceLink
github.comgithub.com/freedomofpress/securedrop-client/blob/8dc8bb6e307b13876d67f72d8a071202e2f39ab5/changelog.md?plain=1#L8
github.comgithub.com/freedomofpress/securedrop-client/commit/e518adaf897e7838467ccf9e1f28152ae6fe3655
github.comgithub.com/freedomofpress/securedrop-client/security/advisories/GHSA-2jrc-x8fq-prvc

Frequently asked

questions

It is conditionally remote: exploitation requires that an attacker has already compromised the SecureDrop Server and can control the gzip archives sent to the Client. From that position, they can indirectly trigger the vulnerability when the Client processes the malicious archive.

The attacker must have sufficient access to the SecureDrop Server to generate or modify the archives the Client consumes. Direct access to the Client workstation is not required, but compromise of the Server—accessible only via Tor and hardened—remains a significant prerequisite.

The safest pattern is to treat all archive filenames as untrusted and enforce strict path confinement: normalize each path, reject absolute or traversal paths, and verify that the target location is within a predefined extraction directory before writing any file.

Preventive measures include using well-reviewed archive libraries that provide safe extraction helpers, standardizing path normalization and base-directory checks, applying least privilege to extraction processes, and conducting focused security reviews for any code that writes files based on remotely supplied paths.

Find out what's already
exploitable in your codebase.