The patch that wasn't: Cisco emits fresh fixes for NTLM hash-spilling vuln and XSS-RCE combo in Jabber app

Wormable nasty still doesn't need any user input to pwn target devices


A previous patch for Cisco's Jabber chat product did not in fact fix four vulnerabilities – including one remote code execution (RCE) flaw that would allow malicious people to hijack targeted devices by sending a carefully crafted message.

Norwegian infosec biz Watchcom spotted the vulnerabilities, having been asked by a client to verify that a previous patch for CVE-2020-26085 worked as advertised. Instead Watchcom found that the September update didn't fix the underlying problems.

A cross-site scripting (XSS) vuln leading to an RCE, CVE-2020-26085 was rated at 9.9 on the 10-point CVSS v3 scale, falling squarely into the "critical" bracket. It was uncovered by Watchcom in June this year and Cisco issued patches on 2 September that allegedly fixed it, as well as three other vulns.

Referring to the XSS-RCE vuln, Watchcom said in a blog post: "This vulnerability does not require user interaction and is wormable, since the payload is delivered via an instant message. This means that it can be used to automatically spread malware without any user interaction."

Watchcom added: "The patch released in September only patched the specific injection points that Watchcom had identified. The underlying issue was not addressed. We were therefore able to find new injection points that could be used to exploit the vulnerabilities."

The critical vuln is accompanied by two others: CVE-2020-27132 and CVE-2020-27127. These are respectively a password hash-stealing information disclosure flaw and a command injection vuln.

Taken together, the three work by allowing the attacker to escape from the Chromium Embedded Framework sandbox, which is the tech underpinning the video 'n' text-based messaging app. Poorly sanitised input handled by the XMPP protocol, which is implemented as Jabber's text component, allows nasty sorts to execute arbitrary text strings in the name attribute of a file tag of a file sent over XMPP.

By abusing this, said Watchcom, it is possible to "inject arbitrary HTML tags into the file transfer message before sending it."

The NTLM password hash can be collected, Watchcom added, by "sending a message that contains a malicious <img> tag, an attacker can force the victim's Cisco Jabber client to interact with a file share the attacker controls."

"If the file share requires authentication, the victim's NTLM password hash will be sent," said the firm's researchers in a statement. If that password happens to be shorter than eight characters, it is currently child's play to recover the plaintext version.

Cisco sent us a statement:

"Following the release of software fixes on September 2, Cisco and a security researcher independently discovered additional, distinct vulnerabilities through testing. We followed our well-established security vulnerability process to address and disclose these vulnerabilities on December 10. Cisco maintains a very open relationship with the security community, and we view this as vital to helping protect our customers’ networks. We thank the security researchers at Watchcom for reporting the vulnerabilities they discovered."

Admins whose orgs use Cisco Jabber should install the latest patches from Cisco, which are available here. ®


Keep Reading

Biting the hand that feeds IT © 1998–2021