The recently declassified U.S. Department of State classification guide describes the Roger channel this way.13
Roger Channel messages are controlled by the Assistant Secretary, INR [Head of State Department Intelligence]. They are used to report sensitive intelligence matters and have very limited distribution. Roger Channel should normally be classified SECRET for a duration of 25 years, or marked 25Xl-human if they reveal the identity of a human intelligence source. (Inclusion of information here on Roger Channel does not constitute authority to initiate messages in this channel. This will normally be done by an OCA [Original Classification Authority].)
Essentially, the Roger channel was a State Department–controlled sensitive communication channel that was intentionally separated from CIA’s classified communication channel in the embassy so that, among other reasons, State’s diplomats could communicate with the State Department in D.C. about CIA without CIA being aware of it. For instance, if any ambassador wanted to complain to his bosses about CIA’s COS, he would do it away from CIA’s prying eyes through the Roger channel.
Returning to the question of how the KGB learned the identities of case officers such as Boagtyr or Augustenborg, one obvious source was the bugged IBM typewriters that typed Roger-channel messages acknowledging the original deployments of these officers.
While acknowledging that information typed on IBM Selectrics might have informed the KGB of CIA case officer identities, Jon LeChevet said the KGB had many other ways of figuring out which embassy officers were legitimate versus intelligence officers:
Spotting station [CIA] personnel was a no-brainer and was often the topic of conversation among State/AID, etc., people. Most of the time the gossip was correct, so it would also be a no-brainer for the KGB to determine who was in the Station using the locals in the Embassy (who handled all travel, car bookings, vacations, cultural events, etc.). The fact that the KGB knew exactly who was an intelligence officer was a given, not a mystery—we were tightly controlled by the local employees who all reported to the KGB. No need to read our traffic to identify IC personnel, it was given to the KGB through the employment of locals in service and support positions.
Based upon LeChevet’s observations (later corroborated by KGB counterspy Yuri Totrov),14 the GUNMAN typewriter implants and Roger-channel responses may have aided the KGB’s efforts to unmask CIA case officers but were only one of many different ways the KGB had of doing this.
But even if the GUNMAN bugs helped identify CIA case officers, these unmaskings did not necessarily tip the KGB off to the identities of those case officers’ assets. Case officers were rigorously trained in countersurveillance (e.g., shaking off tails), covert communications, and many other types of tradecraft, such as dead drops and brush passes, and were held accountable for scrupulously protecting the identities of their assets. Thus, with normal tradecraft, case officers rarely made mistakes that informed the opposition who their assets were.
With normal tradecraft.
But in 1984, a KGB officer and CIA asset named Sergei Vorontsov told CIA that the KGB had for years been covertly sprinkling “spy dust” on CIA case officers (probably through FSNs under KGB control in the embassy) that allowed the KGB to not only track case officers after they slipped their KGB tails but to identify Soviet citizens who came in contact with the case officers.15 The chemical nitrophenyl pentadienal was invisible but glowed under ultraviolet (UV) light in very small quantities and normally transferred from one person to another through contact or exchange of objects (such as microfilm containers and cameras).16 Thus, even with flawless tradecraft, CIA case officers, identified through GUNMAN bugs in typewriters—or the other sources LeChevet referred to—could have led the KGB straight to their assets.
Taking all these factors into consideration, it seems likely that the GUNMAN typewriter implants were somewhat helpful—although far from necessary—to the KGB in unmasking CIA case officers and that it’s further possible, although far from certain, that some of these unmaskings led the KGB to some CIA human assets through spy dust or other means.
Presented with the weak, circumstantial evidence linking GUNMAN implants to human asset roll-ups in Moscow, Gandy said in 2018, “Well, Hathaway told me about other roll-ups [excluding Ogorodnik, Filatov, Nilov, and Kapustin] that Hathaway worried were lost through text [typing], although before Project GUNMAN, I thought we were losing text through MUTS. Remember, I told Hathaway from the very start I thought the problem was text, and he did invite me over there three separate times [twice in 1978 and once in 1981], and his main concern all along was HUMINT losses.”
One final piece of evidence that indicates how much damage the GUNMAN implants did to U.S. national security was the very existence of the GUNMAN system itself. Although the Soviets put a high priority on intelligence gathering, Russia was (and is) a poor country that was not in the habit of focusing scarce resources on unproductive pursuits.
The typewriter implant itself cost a lot to design and manufacture, especially with the eighteen different ultrasophisticated hides that NSA uncovered. GUNMAN implants also required staffing at listening posts, maintenance, analysis, and other activities that cost rubles. Gandy estimated that over the eight years GUNMAN was known to operate, the KGB operated six listening posts around the chancery, each with a crew of three officers.
Moreover, the KGB invested in four technical upgrades to the system after its original introduction in 1976. Bottom line: the Soviets believed the GUNMAN system was worth the investment over an eight-year period. Would they have felt this way if reading typewriters in the U.S. embassy didn’t compromise U.S. national security in a meaningful way?
So, how much damage did the GUNMAN implants do, either on compromised arms negotiations or human intelligence gathering in Moscow?
Either a little or a lot, depending upon whom you choose to believe.
13. Lessons About the Russians for Today
Annapolis, Maryland, Early Spring, 2018
I sat in Charles and Freda Gandy’s spacious house overlooking the water, marveling at the spectacular view and sipping the second Diet Coke Gandy had given me. I had been running draft chapters of this book by Gandy for months and often stopped by his house to discuss the drafts. I also frequently shared source material for the book with Gandy, especially Russian-language accounts of the embassy typewriter bugging and microwave attacks that I had translated for his benefit.
Gandy shook his head, a sad look playing on his usually cheery face as he read one of these translations, a 2005 article, “Moscow’s Sensitive Ears,” in Independent Gazet by Dimitry Prokhorov, who has written extensively on Russian and Israeli “special services.”
In 1984 masters from the 16 Department [of KGB] mounted bugs in 30 new typewriters intended for the US Embassy in Moscow and US Consulate General in Leningrad.1
“We only found sixteen,” Gandy said while he read and reread the translated document. “That means we missed fourteen or that there was a fifth generation that went undetected.”
“How was that possible?” I asked.
“Don’t know. Perhaps the other side knew we were doing a swap in ’84 and held some back so that we’d think we got them all, or maybe they slowly slipped them in one at a time after the swap. They [KGB] were sophisticated like that, and patient. They also had free run of the chancery until after the Marine guard scandal in ’87 and endless opportunities to put in new machines, or do swaps during maintenance.” (Several Marine guards at the Moscow embassy had confessed to being seduced by KGB honey traps employed at the embassy, potentially compromising secure areas and leading to congressional investigations and the ultimate expulsion of all but a few FSNs from the Moscow embassy.)
“So it might be true? There really could have been another fourteen?”
A faraway look came over Gandy as he looked out at the large sailboat at his backyard dock, still protectively wrapped for the winter. “Could be. I’m sitting here kicking myself for not thinking of
it. I was so busy doing other things at the time and let others take over. After GUNMAN, Deeley’s organization was renamed INFOSEC [information security] from COMSEC [communications security] to expand NSA’s charter from com links to things like computers and typewriters, and NSA was given much more responsibility over there [for embassy security in Moscow]. But I moved on.” Gandy sighed deeply.
“So after all that Sturm und Drang with CIA and State, after your tortuous six-year odyssey, the cease-and-desist from DCI, Deeley, Reagan, Arneson … all of it, the KGB might have just kept going?”
Gandy said, “Well, maybe those extra fourteen [undiscovered GUNMAN implants] mattered, and maybe they didn’t. I do think we put at least a temporary dent in their operation, because right after we pulled the typewriters in ’84, they dug trenches all around the compound and buried coax cables to improve the sensitivity of MUTS [microwave attacks on the embassy] so that they could gather text, and maybe voice, too, that way. They wouldn’t have gone to all that extra trouble with the trenches, I think, unless they’d lost some important access.”
“Okay,” I said slowly, “but what you’re telling me is that even if the Russian journalist is wrong about the extra fourteen GUNMAN machines, the KGB made up for the loss of GUNMAN by beefing up MUTS.”
“Well, sure,” Gandy said, brightening. “They never stop.”
As I digested that, I ran the phrase they never stop over and over in my head. I said, “So it’s a safe bet they’re going strong against us right up till today.”
Gandy smiled as he used a phrase he’d picked up from one of his grandkids. “Well, duhhh.”
“Okay,” I said, feeling sheepish for asking such a dumb question. “Does GUNMAN teach us anything useful for today about the Russians? Does it have some value in 2018?”
“I’m not sure I follow you,” Gandy answered.
“Well,” I said, “did GUNMAN play a role in getting us to where we are today with the Russians? For instance, did we teach them back then how to treat us? Think about it. They leaned way forward, took a huge risk invading our embassy, and in the end, what did we do when we finally caught them? We took six years, all the while denying, finger-pointing, and fighting each other. Ultimately, we did nothing to the Russians to hold them accountable for what they did to us. Sounds a whole lot like the 2016 election and its aftermath to me.”
Gandy leaned forward in his stuffed chair and said, “You’d have to go much further back than GUNMAN for the ‘we taught them how to treat us’ thing. They bugged the Great Seal in the ’50s, and we found a hundred microphones in the embassy in the ’60s, and all that time, up to the present as far as I know, they’ve been hammering us with microwaves. And throughout all of that, they [Russia] suffered no consequences.”
I said, “So what we’re seeing today with the hacking, election meddling is business as usual. If you’re the Russians, as long as you stay away from kinetics [overt military warfare] and invasions … take your best shot at America, and nothing bad will happen to you.”
“Yeah. I don’t imagine they are too worried about sanctions.”
As Gandy spoke, I thought back to his who-hates-whom chart. “It occurs to me,” I said, “that we have very little influence over the Russians, but we do have some control, at least in theory, over ourselves.”
Gandy said, “Go on.”
I opened the printed draft of the book to the who-hates-whom chart. “We’ve got one of these right now, you know. Different boxes—different players, maybe—but not a whole lot different. CIA and DOD still despise each other. Ditto for FBI and CIA; State still doesn’t like CIA; and the White House is in conflict with plenty of key players. The only new box for your chart is ODNI [Office of the Director of National Intelligence that oversees all seventeen U.S. intelligence agencies], and I can tell you from firsthand experience as a DNI guy, everyone hated us as meddling, micromanaging, incompetent bean counters.”
Gandy chuckled.
“So it seems to me that now, as then, the Russians may instigate bad things, but we inflict most of the damage on ourselves by fighting each other after the Russians do to us what the Russians always do.”
“I can’t argue with a single thing you said,” Gandy answered, “but what do you propose to do about it?”
I answered, “That’s a tough one. When you think about it, three of the key actors in GUNMAN—you, Hathaway, and LeChevet—were not turf defenders who battled other agencies on principle. You three did cooperate—a lot—to get to the truth, but it wasn’t nearly enough to overcome the enmity of your respective organizations. LeChevet, for example, was stepped on hard by his masters for helping you, and you were squashed by CIA. If it weren’t for Walt Deeley saying, “Screw everyone, we’re going to solve this,” and going directly to the president, the typewriter bugs would never have been found.”
Gandy got up to get me another Diet Coke. On the way back from the refrigerator, he said, “Yeah, Walt was the real hero of GUNMAN. But there sure aren’t many like him.”
I pondered that. If the U.S. intelligence community couldn’t overcome its internal friction without a Walt Deeley—someone willing to risk everything to do the right thing—then the future didn’t look so rosy because leaders like Walt Deeley’s only came along once in a generation. Or a century.
I stared at the who-hates-whom chart, mentally superimposing new names, boxes, and links into it to bring it up to date. I tried to imagine the current set of actors—or indeed any set of actors past, present, or future—putting aside politics, turf wars, and tribal Washington hatreds to unite against the Russian threat. The 2016 election episode taught the Russians that, when they get caught conducting active measures to destabilize America, America wouldn’t do much to retaliate but would spend years tearing itself apart.
Considering this, I wondered how the current inhabitants of the who-hates-whom chart would react when Russia, encouraged by 2016, tested America further through a crippling cyberattack on U.S. banks, the stock exchange, or even the power grid. The Russians would deny the attack, of course, knowing that some of the players on the modern who-hates-whom chart would accept their denial, thereby spawning a protracted conflict with those on the chart who wanted to retaliate against the Russians.
For the Russians, it would all be good. The damage to the United States from the cyberattack would be nice, but the recriminations, finger-pointing, and bitter political disputes that would follow the attack would be far, far better. Going all the way back to Lenin, dividing Russia’s enemies has not just been a means to an end for the Kremlin but a desired end all unto itself.
At length, I said to Gandy, “Boating season will be here soon, right?”
“Yep.”
“Well,” I said, putting away the chart and standing up, working the kinks out of my legs, “let me help you get the boat ready and rigged for spring. Could be stormy weather ahead.”
Author’s Note
This work was derived from declassified U.S. government documents, interviews, press accounts, books on the period, and publicly available technical documents, mostly from Russian-language sources, on technical surveillance. No classified information was used in the book, and all documents marked SECRET or TOP SECRET in the book have since been declassified. Neither NSA nor CIA, who approved release of the material as unclassified, acknowledges the accuracy of any of the accounts.
Nothing in this work represents validation, based upon my service as a U.S. intelligence official, of any of the claims described in the book from public literature—such as Russian-language literature on technical surveillance techniques—regarding the technical viability of techniques described or their use by any U.S. government entity, including NSA. For instance, possible ways that Gandy may have gained confidence in 1978 that Russian intelligence was reading text from machines in the U.S. embassy were speculation on my part and not based upon any information from either Gandy or from any U.S. government sources.
The conversations descr
ibed here were based upon the recollections of at least one participant in each of the conversations, and multiple participants, where possible.
In searching for completely unclassified sources for technical surveillance techniques discussed in the book, I made an intriguing and, to me, disturbing discovery. Techniques such as radar flooding (microwave attacks) and microphonic exploitation (recovering voice signals from vibration of electronic circuits) are widely described in entry-level Russian-language textbooks on information security under the heading of “leakage of information through technical channels.” Information security curricula at Russian universities similarly cover “leakage of information through technical channels” in depth in introductory courses.
In contrast, although TSCM tradecraft—and techniques such as RF imposition and microphonics—are known in the West, U.S. texts and courses on information security emphasize computers, software operating systems, and networks, but rarely make more than passing mention of the physics of the electronics that underlie information systems, which the Russians exploit through technical channels. In essence, the Russians look at computers and networks not just as conveyors of digital bits but also as radio transmitters and receivers, magnetic flux generators and magnetometers, optical emitters and shot sensors, acoustic speakers and microphones.
Computers and networks were not designed to exhibit these physical characteristics, but Maxwell’s equations, which describe electromagnetic fields associated with electric currents, dictate that they do have these characteristics. And the Russians view Maxwell’s equations as offering up many opportunities both to peek and poke at computers, communications systems, and networks. The MUTS microwave attacks on the U.S. embassy (which continued at least until the early 2000s, according to a retired State Department security specialist) are an example of such peeking and poking.
The Spy in Moscow Station Page 24