The massive data breach at the U.S. Office of Personnel Management (OPM) that exposed background investigations and fingerprint data on millions of Americans was the result of a cascading series of cybersecurity blunders from the agency’s senior leadership on down to the outdated technology used to secure the sensitive data, according to a lengthy report released today by a key government oversight panel.
The 241-page analysis, commissioned by the U.S. House Oversight & Government Reform Committee, blames OPM for jeopardizing U.S. national security for more than a generation.
The report offers perhaps the most exhaustive accounting and timeline of the breach since it was first publicly disclosed in mid-2015. According to the document, the lax state of OPM’s information security left the agency’s information systems exposed for any experienced hacker to infiltrate and compromise.
“The agency’s senior leadership failed to fully comprehend the extent of the compromise, allowing the hackers to remove manuals and other sensitive materials that essentially provided a roadmap to the OPM IT environment and key users for potential compromise,” the report charges.
Probably the most incisive portion of the assessment is the timeline of major events in the breach, which details a series of miscalculations on the part of the OPM leadership. The analysis paints the picture of a chronic — almost willful — underestimation by senior leadership at OPM about the seriousness of the threat facing the agency, until it was too late.
According to the report, the OPM first learned something was amiss on March 20, 2014, when the US-CERT notified the agency of data being exfiltrated from its network. In the ensuing weeks, OPM worked with US-CERT to implement a strategy to monitor the attackers’ movements to gather counterintelligence.
The only problem with this plan, according to the panel, was that the agency erroneously believed it had cornered the intruder. However, the hacker that OPM and US-CERT had eyes on wasn’t alone. While OPM monitored the first hacker [referred to in the report only as Hacker X1] on May 7, 2014 another hacker posed as an employee of an OPM contractor (Keypoint) performing background investigations. That intruder, referred to as Hacker X2, used the contractor’s OPM credentials to log into the OPM system, install malware and create a backdoor to the network.
As the agency monitored Hacker X1’s movements through the network, the committee found, it noticed hacker X1 was getting dangerously close to the security clearance background information. OPM, in conjunction with DHS, quickly developed a plan to kick Hacker X1 out of its system. It termed this remediation “the Big Bang.” At the time, the agency was confident the planned remediation effort on May 27, 2014 eliminated Hacker X1’s foothold on their systems.
The decision to execute the Big Bang plan was made after OPM observed the attacker load keystroke logging malware onto the workstations of several database administrators, the panel found.
“But Hacker X2, who had successfully established a foothold on OPM’s systems and had not been detected due to gaps in OPM’s security posture, remained in OPM’s systems post-Big Bang,” the report notes.
On June 5, malware was successfully installed on a KeyPoint Web server. After that, X2 moved around OPM’s system until July 29, 2014, when the intruders registered opmlearning.org — a domain the attackers used as a command-and-control center to manage their malware operations.
Beginning in July through August 2014, the Hacker X2 exfiltrated the security clearance background investigation files. Then in December 2014, 4.2 million personnel records were exfiltrated.
On March 3, 2015, wdc-news-post[dot]com was registered by the attackers, who used it as a command-and-control network. On March 26, 2015, the intruders begin stealing fingerprint data.
The committee found that had the OPM implemented basic, required security controls and more expeditiously deployed cutting edge security tools when they first learned hackers were targeting such sensitive data, they could have significantly delayed, potentially prevented, or significantly mitigated the theft.
For example, “OPM’s adoption of two-factor authentication for remote logons in early 2015, which had long been required of federal agencies, would have precluded continued access by the intruder into the OPM network,” the panel concluded.
Unfortunately, the exact details on how and when the attackers gained entry and established a persistent presence in OPM’s network are not entirely clear, the committee charges.
“This is in large part due to sloppy cyber hygiene and inadequate security technologies that left OPM with reduced visibility into the traffic on its systems,” the report notes. “The data breach by Hacker X1 in 2014 should have sounded a high level, multi-agency national security alarm that a sophisticated, persistent actor was seeking to access OPM’s highest-value data. It wasn’t until April 15, 2015 that the OPM identified the first indicator that its systems were compromised by Hacker X2.”
The information stolen in the breach included detailed files and personal background reports on more than 21.5 million individuals, and fingerprint data on 5.6 million of these individuals. Those security clearance background reports often included extremely sensitive information, such as whether applicants had consulted with a health care professional regarding an emotional or mental health condition; illegally used any drugs or controlled substances; experienced financial problems due to gambling.
The intrusion, widely attributed to hackers working with the Chinese government, likely pointed out which federal employees working for the U.S. State Department were actually spies trained by the U.S. Central Intelligence Agency. That’s because — unlike most federal agencies — the CIA conducted its own background checks on potential employees, and did not manage the process through the OPM.
As The Washington Post pointed out in September 2015, the CIA ended up pulling a number of officers from its embassy in Beijing in the wake of the OPM breach, mainly because the data leaked in the intrusion would have let the Chinese government work out which State Department employees stationed there were not listed in the background check data stolen from the OPM.
As bad and as total as the OPM breach has been, it’s remarkable how few security experts I’ve heard raise the issue of what might be at stake if the OPM plunderers had not simply stolen data, but also manipulated it.
Not long after congressional hearings began on the OPM breach, I heard from a source in the U.S. intelligence community who wondered why nobody was asking this question: If the attackers could steal all of this sensitive data and go undetected for so long, could they not also have granted security clearances to people who not only didn’t actually warrant them, but who might have been recruited in advance to work for the attackers? To this date, I’ve not heard a good answer to this question.
A copy of the 110 mb report is available here (PDF).