//
OFFLINE — VIEWING CACHED CONTENT
← Back to Archive

IC BRIEF

Current as of 1712 EDT (UTC-04), Sunday 26 April 2026

Contents

12 stories from 20 sources across 15 organizations

BOTTOM LINE UP FRONT

We assess that Russian and Iranian intelligence services are very likely to sustain aggressive cyber and HUMINT operations against Western targets through mid-2026, with at least one additional major espionage case or campaign attribution becoming public within 60 days. This is a high-confidence judgment grounded in concurrent documented operations: Germany's formal espionage investigation into Russia's 300-plus-account Signal phishing campaign, Iran's concurrent protester executions and Islamic Revolutionary Guard Corps (IRGC) mass arrests across six provinces, and CyberAv3ngers' active exploitation of U.S. critical infrastructure PLCs under the joint AA26-097A advisory.

The U.S. intelligence community's defensive posture is not keeping pace. Cybersecurity and Infrastructure Security Agency (CISA)'s exclusion from Anthropic's Mythos frontier AI evaluation, while NSA and 40-plus private entities hold access, exposes a capability gap likely to draw sustained congressional pressure through summer 2026 as Section 702 reauthorization and House Permanent Select Committee on Intelligence (HPSCI)'s China technology oversight converge on the same calendar. The indicator that would alter this assessment: CISA gaining formal frontier-model access, or a diplomatic pause in either theater that delays public attributions.

IC Technology & Surveillance

CISA Last in Line for Access to Anthropic Mythos AI System

CISA has no access to Anthropic's Mythos Preview model despite its role as the nation's lead cyber defense agency. Entities with access include the NSA, Commerce Department's Center for AI Standards and Innovation, and 40-plus companies and organizations via Anthropic's Project Glasswing initiative. Bloomberg separately reported that an unauthorized private Discord group focused on tracking unreleased AI models gained access and has been using Mythos regularly. Anthropic restricted Mythos due to concerns about its ability to rapidly identify and exploit software vulnerabilities.

Analyst Note: Granting a restricted model to NSA and 40-plus private-sector entities while excluding the statutory cyber defense agency inverts the expected governance hierarchy. The Discord disclosure sharpens the problem: per Bloomberg, an unauthorized community tracking unreleased AI has been using Mythos regularly, and if an ad hoc group can reach a model flagged for rapid exploit-finding, the access-control regime around frontier AI is weaker than Anthropic's public posture suggests. CISA's guidance on AI-discovered vulnerabilities will likely remain compromised in the coming months absent a formal access agreement with Anthropic, though classified channels or NSA sharing arrangements outside public reporting could make the gap narrower than the secondary rewrites from CSO Online and Computerworld indicate. Congressional engagement is visible in the House Homeland Security Committee's jailbroken-AI briefing, but engagement is not access.

Sources:

Researchers Uncover Pre-Stuxnet fast16 Malware Targeting Engineering Software

SentinelLabs researchers Kamluk and Guerrero-Saade published findings on the fast16 framework consisting of three components: svcmgmt.exe carrier with embedded Lua 5.0 VM, fast16.sys kernel driver compiled July 19 2005 with 101 configuration rules for intercepting and modifying executable code during disk reads, and svcmgmt.dll reporting module using a named pipe. The driver targets binaries compiled with Intel C/C++ compiler and corrupts floating-point unit routines. Three engineering suites were targeted: LS-DYNA 970 for crash analysis, PKPM for Chinese civil engineering, and MOHID for hydrodynamic modeling. The fast16 name appeared in the Shadow Brokers' April 2017 leak within the NSA's Territorial Dispute deconfliction document with the annotation: Nothing to see here, carry on. This predates Stuxnet by at least five years.

Analyst Note: The July 2005 compilation timestamp and the NSA Territorial Dispute annotation, "nothing to see here, carry on," establish fast16 as the earliest publicly documented IC framework targeting engineering-simulation software, five years before Stuxnet. Per SentinelLabs' Kamluk and Guerrero-Saade, with technical artifacts independently verifiable against the 2017 Shadow Brokers' leak, the selection of LS-DYNA, PKPM, and MOHID, crash, civil-engineering, and hydrodynamic modeling tools, points toward intelligence collection against adversary industrial capacity rather than sabotage, consistent with SIGINT operations against weapons-program modeling. PKPM's specificity to Chinese civil engineering could reflect a non-U.S. origin, though the Territorial Dispute annotation's framing implies NSA awareness of the tool as a friendly or tolerated operation. Additional public analysis is likely within the coming weeks given the publication's depth and the Territorial Dispute thread. Researchers applying fast16 indicators to the broader Shadow Brokers corpus will likely surface additional IC-linked malware from the same era.

Sources:

Scattered Spider Co-Conspirator Pleads Guilty in Cybercrime Case

Tyler Robert Buchanan, 24, of Dundee, Scotland, pleaded guilty on April 17 in the Central District of California to one count of conspiracy to commit wire fraud and one count of aggravated identity theft, facing a statutory maximum of 22 years. Operating under the handle Tylerb, Buchanan admitted conspiring with Scattered Spider members to launch tens of thousands of SMS phishing attacks in 2022 leading to intrusions at Twilio, LastPass, DoorDash, Mailchimp, and at least a dozen other technology companies. He personally stole at least 8 million dollars in cryptocurrency. Spanish authorities reportedly seized 27 million dollars at his arrest at Palma airport in June 2024. Sentencing is set for August 21, 2026. Co-defendant Noah Michael Urban is already serving 10 years.

Analyst Note: The Buchanan plea, the second major Scattered Spider conviction, consolidates the DOJ's cooperation-driven prosecution strategy against The Com's social-engineering infrastructure. The DOJ press release is the authoritative primary source; CSO Online is a secondary rewrite. The 22-year statutory ceiling combined with a mandatory consecutive 2-year aggravated identity theft minimum and Urban's 10-year precedent makes a sentence of 10 or more years at the August 21 sentencing likely. The DOJ filing newly confirms $27 million seized at Palma airport, establishing Spanish-U.S. coordination that prior reporting had not detailed. A cooperation agreement, publicly unconfirmed, could pull Buchanan's sentence well below Urban's benchmark if he provides intelligence on other Com members.

Sources:

Prior Reporting

Space Force Awards $3.2 Billion for Golden Dome Space-Based Interceptor Prototypes

Space Systems Command announced on April 24 the award of 20 Other Transaction Authority agreements worth up to 3.2 billion dollars combined across 12 companies: Anduril Industries, Booz Allen Hamilton, General Dynamics Mission Systems, GITAI USA, Lockheed Martin, Northrop Grumman, Quindar, Raytheon, Sci-Tec, SpaceX, True Anomaly, and Turion Space Corp. Companies will demonstrate space-based interceptor prototypes by 2028, targeting a proliferated Low Earth Orbit (LEO) constellation carrying kinetic interceptors capable of destroying missiles during boost, midcourse, and glide phases. Col. Bryon McClain, Program Executive Officer (PEO) for Space Combat Power, stated that adversary capabilities are advancing rapidly. The overall Golden Dome program is estimated at 185 billion dollars with expert projections ranging to 3.6 trillion depending on interceptor fleet size.

Analyst Note: The awards launch the most ambitious space-based missile defense effort since Strategic Defense Initiative (SDI), but the 2028 prototype target is unlikely: no nation has demonstrated kinetic intercept from LEO, and historical space-weapons programs have run 15 or more years from contract to initial operating capability. The Other Transaction Authority (OTA)'s parallel-prototype structure may be a deliberate hedge, buying down technical risk rather than committing to a timeline Space Systems Command expects to meet. A proliferated LEO interceptor constellation will generate adversary intelligence requirements across Russian and Chinese space-awareness programs, with technical collection on adversary countermeasures likely becoming a priority within the next 12 months. The mix of non-traditional firms alongside established primes reflects a deliberate effort to diversify the interceptor industrial base.

Sources:

FBI CISA NSA Warn Iranian Threat Actors Actively Exploiting US Critical Infrastructure PLCs

Joint advisory AA26-097A was published April 7 by six agencies: FBI, CISA, NSA, EPA, DOE, and US Cyber Command. Threat actors are linked to Iran's IRGC Cyber Electronic Command operating under the CyberAv3ngers persona. Since at least March 2026, actors have exploited internet-exposed Rockwell Automation and Allen-Bradley PLCs, specifically CompactLogix and Micro850 controllers, plus Siemens S7 and Modbus-compatible devices across Government Services, Water and Wastewater Systems, and Energy sectors. The attack chain involves initial access via internet-exposed PLCs, deployment of Dropbear Secure Shell (SSH) on port 22 for persistent Command and Control (C2), extraction and modification of Programmable Logic Controller (PLC) project files, and Supervisory Control and Data Acquisition (SCADA) Human-Machine Interface (HMI) display manipulation where operators see false readings while malicious control logic executes.

Analyst Note: The false-operator-display technique, where operators see normal readings while malicious control logic executes, marks a capability escalation from the December 2023 Aliquippa compromise, where simpler Unitronics default-credential exploitation left no comparable detection gap. Per a single Picus Security analysis of the AA26-097A advisory, detection signatures now exist for Dropbear SSH persistence, PLC project file extraction, and SCADA HMI manipulation, techniques the six-agency advisory described only at the advisory level without defender-actionable specificity. Whether CyberAv3ngers escalate from access maintenance to destructive attack within 90 days is uncertain, given demonstrated multi-platform capability, active exploitation since at least March 2026, and ongoing U.S.-Iran military conflict. The sustained access may instead reflect long-term intelligence collection against infrastructure, consistent with Chinese APT operations against comparable targets where access is held for years without escalation, though the wartime context reduces that precedent's force. Alongside other Iran reporting from this cycle, the IRGC's concurrent internal security operations and external cyber activity illustrate the breadth of Iranian intelligence operations under wartime conditions.

Sources:

Prior Reporting
Watch Items - CISA Mythos access: Anthropic may expand Project Glasswing to include federal cyber defenders. The unauthorized Discord group accessing Mythos is a potential security incident. - CyberAv3ngers' active exploitation of US PLCs: AA26-097A documents ongoing attacks since March 2026. The Iran conflict increases the risk of escalation, though whether a destructive ICS attack occurs within 90 days is uncertain. - Golden Dome space-based interceptor prototypes: 12 companies awarded 3.2 billion in OTAs. Integration timeline of 2028 is ambitious.

Adversary Intelligence

Iran Executes Alleged Israeli Agent Over Sabotage During Nationwide Protests

Erfan Kiani was hanged early Saturday April 25 in Isfahan after Iran's Supreme Court upheld his death sentence. Per Tasnim News Agency, Kiani was charged with destroying public and private property, arson, carrying Molotov cocktails, carrying cold weapons, blocking vehicle paths, and attacking security agents during the January 2026 protests. The regime labeled him a hired thug of Mossad. Iran Human Rights reports Kiani is the ninth protester executed in connection with the December 2025 and January 2026 protests. Human rights organizations have flagged the absence of transparent trial information and concerns about access to legal counsel.

Analyst Note: The espionage label applied to Kiani, designated a Mossad-hired thug with no verifiable evidence of Israeli operational ties presented in any source, including Israeli outlets relaying Tasnim claims, continues a post-2022 pattern of reframing domestic dissent as foreign-directed subversion. The same-day IRGC sweep of 240-plus individuals in western provinces signals a coordinated internal security posture: simultaneous prosecution of alleged foreign-linked operatives and mass suppression of domestic unrest. Nine completed executions and additional death sentences still moving through the judicial system make further executions of protest participants very likely through 2026. Kiani may nonetheless have had genuine separatist or foreign-linked activity that Iran's summary proceedings failed to transparently establish: a due-process failure rather than a purely fabricated charge. Iran Human Rights, the primary monitoring body with strong verification methodology, is the authoritative source on the execution count. Israeli and Western reporting relays Tasnim state media without independent corroboration of the espionage designation.

Sources:

Iran Arrests Alleged Veteran Mossad Spy and 15 Separatist Operatives in Nationwide Sweep

On April 25, Iran's Ministry of Intelligence announced the arrest of a veteran Mossad operative in Ilam Province who collected intelligence for Israel and later collaborated with monarchist groups, along with 15 separatist operatives across multiple provinces. Separately, the IRGC Intelligence Organization announced the arrest of over 240 individuals in Kermanshah and Kurdistan provinces, including 155 agents connected to anti-revolutionary groups, 4 suspected Mossad-linked spies, and 7 separatist elements engaged in Improvised Explosive Device (IED) manufacturing. Weapons seized included eight RPG launchers and 2,000 rounds of ammunition. Provinces affected include Ilam, Sistan-Baluchestan, Gilan where a three-member cell with a homemade mortar launcher was found, and Lorestan where a cell receiving cryptocurrency payments was constructing bombs targeting governors' offices.

Analyst Note: The sweep, per Iranian state media with no independent confirmation, is the largest publicly reported mass arrest in Iran's western provinces since the January 2026 protest crackdown, with weapons seizures and a cryptocurrency-funded bomb-making cell in Lorestan pointing to genuine armed-group infrastructure while the IRGC frame blankets verified threats and political dissent alike under foreign-directed subversion. The operation marks a sharp escalation from the April 2 Semnan arrests, targeted counterespionage against three Mossad-recruited operatives, to a 240-plus dragnet across six provinces conflating armed separatists with political opposition. The arrests are likely a force-protection measure in the western border provinces during the ongoing U.S.-Iran military conflict, aimed at suppressing rear-area threats in the coming weeks. The scale is consistent with IRGC deterrent inflation following protest cycles, where claimed figures have repeatedly proved unverifiable.

Sources:

Prior Reporting

German Prosecutors Open Formal Espionage Case as Signal Phishing Campaign Compromises 300+ Political Accounts

German federal prosecutors opened a formal espionage investigation on April 24 based on initial suspicion of espionage. Der Spiegel reported at least 300 accounts belonging to political figures were compromised. Targets include the Bundestag speaker, a senior Christian Democratic Union (CDU) member, civil servants, diplomats, military personnel, and journalists. The attack method involves spoofed Signal support messages requesting PINs, links, or QR codes granting attackers access to messages, chat groups, photos, and files, plus impersonation capability. The German government formally attributed the campaign to Russia. MP Konstantin von Notz stated no one can say with any certainty whether the integrity of MPs' communications is still guaranteed. Moscow denies involvement.

Analyst Note: The attack methodology exploits Signal's device-linking protocol rather than its encryption: spoofed support messages solicit PINs and QR codes that link attacker-controlled devices, preserving cryptographic integrity while bypassing it operationally. Converging reporting anchored by Der Spiegel's 300-plus compromised accounts across parliamentarians, diplomats, military, and journalists elevated this from targeted phishing to one of the most extensive Russian intelligence operations against a NATO parliament disclosed in recent years. Prosecutors' April 24 escalation to formal espionage investigation, beyond the cybersecurity-incident framing, formalized that assessment. Formal attribution to a specific Russian service is unlikely within 30 days: forensic material is substantial, but German prosecutors will likely prioritize the investigation over public attribution. Von Notz's acknowledgment that communications integrity can no longer be guaranteed makes Bundestag restrictions on Signal desktop on parliamentary systems likely, with downstream implications for secure legislative communications across NATO. The campaign's breadth may instead indicate an automated or semi-automated operation rather than bespoke intelligence collection, which would lower assessments of operational sophistication and narrow the probable perpetrator set.

Sources:

Prior Reporting
Watch Items - German espionage investigation into Russian Signal phishing: watch for attribution to a specific service (GRU, SVR, or FSB). 300+ compromised accounts provide substantial forensic material; formal attribution within 30 days is uncertain. - Iran mass arrests: 240+ detained in IRGC sweep including alleged Mossad-linked operatives. Watch for due process concerns and potential use of arrests to justify crackdown on protest movements. - Iran execution of Erfan Kiani: ninth protester executed from December-January protests. Watch for international response and whether Trump raises executions in negotiation context.

IC Oversight & Authorities

House Homeland Security Committee Gets Demo of Jailbroken AI Capabilities

The House Homeland Security Committee's Subcommittee on Counterterrorism and Intelligence hosted a closed-door bipartisan briefing by DHS's National Counterterrorism Innovation, Technology, and Education Center (NCITE) demonstrating jailbroken AI capabilities. Researchers compared censored models including Anthropic's Claude and OpenAI's ChatGPT against abliterated versions with deactivated refusal mechanisms. The demo asked both types to create a plan to attack the America 250 celebration to harm as many attendees as possible. Censored models refused while abliterated models provided step-by-step attack instructions. Chairman Garbarino said he asked a model how to kidnap a member of Congress and it responded in under three seconds. Rep. Ogles noted most of the AI tools are readily off-the-shelf and easy to access.

Analyst Note: Per the subcommittee's own press release, with no independent verification of demonstration methodology, NCITE's side-by-side comparison of censored and abliterated models is the first disclosed briefing of safety-deactivated AI capabilities to a congressional intelligence panel, giving the subcommittee a concrete threat baseline to drive distribution-restriction advocacy. The specific examples, step-by-step attack planning for the America 250 celebration and a kidnapping query returned in under three seconds, anchor the argument on open-weight models where refusal-mechanism removal is trivial. Legislative or regulatory action on dual-use AI distribution controls is likely within the coming months, though scope and enforcement mechanisms remain undefined. The demonstration may have overstated marginal risk by pairing worst-case abliterated outputs against best-case censored refusals, understating how much comparable harm-enabling content already circulates through non-AI channels. The CISA-Anthropic Mythos access gap, with federal cyber defenders excluded from frontier model evaluation while Congress presses for restrictions, exposes the structural disconnect any regulatory framework will need to close.

Sources:

HPSCI Chairs Crawford and Mast Issue Joint Statement on China Tech Competition Briefing

House Permanent Select Committee on Intelligence Chairmen Crawford and Mast released a joint statement on April 21 following a classified briefing on technology competition with China. The briefing addressed the scope of Chinese efforts to acquire advanced technology including AI and semiconductors through both legal and illicit channels, and the intelligence community's assessment of the competitive landscape. The CCP Select Committee separately requested a Commerce Department investigation into CCP-tied technology threats, naming the same technology acquisition vectors the HPSCI briefing addressed.

Analyst Note: The Crawford-Mast joint statement, sourced solely to committee press releases with no independent verification of the IC assessment's substance, converts an internal oversight event into a cross-committee policy signal, with the CCP Select Committee's concurrent Commerce Department investigation request indicating HPSCI is likely to pursue export-control tightening through legislative language in the coming months rather than through IC-internal collection guidance. The shift is the cross-committee coordination itself, not the briefing, converting a classified oversight routine into a legislative action signal. The joint statement may instead represent standard post-briefing messaging, and Commerce Department investigations of this type frequently stall at the preliminary-inquiry stage. Whether the IC assessment presented supported the committees' interventionist framing or offered a more nuanced competitive picture remains unreported.

Sources:

Prior Reporting
Watch Items - Section 702 expires April 30: watch for another stopgap or brief authority lapse. Brennan Center disclosed new FISA Court opinion revealing continuing FBI violations. - CIRCIA final rule expected May 2026: defines mandatory incident reporting requirements for critical infrastructure operators. - HPSCI-HFAC cross-committee coordination on China tech competition: watch for Commerce Department's response to CCP-tied tech investigation request.

Counterintelligence & Tradecraft

CIA Ran MK-ULTRA Experiments on Prisoners of War, Declassified Documents Confirm

The Intercept published an investigation on April 26 drawing on the National Security Archive's collection of over 1,200 declassified CIA records released between December 2024 and April 2025. Key documents include an April 5, 1950 memo to CIA Director Hillenkoetter outlining Project Bluebird's goals and budget, and an August 8, 1951 account confirming 25 North Korean POWs as experimental subjects for polygraph testing, drug combinations including sodium amytal with benzedrine and picrotoxin, and hypnosis. A 1952 memo to Allen Dulles stated the CIA proceeded despite finding no evidence any American POWs were subjected to drug-induced hypnosis. The collection is accessible via the Digital National Security Archive through ProQuest.

Analyst Note: The 1,200-document Digital National Security Archive (DNSA) release is the most substantial primary-source disclosure on MK-ULTRA's POW experimentation track since the 1977 Senate hearings. Per a single Intercept investigation drawing on records independently verifiable through ProQuest, the April 1950 Bluebird memo and 1951 POW account together confirm what those hearings could only infer from fragments: CIA proceeded despite finding no evidence that adversaries had applied the same techniques. Additional reporting from the archive is likely in the coming months given its scale and public accessibility, though the collection may already have been substantially mined by researchers before this publication, limiting the prospect of further revelations. Congressional oversight action is unlikely this quarter absent sustained media pressure.

Sources:

Watch Items - Watch for additional declassified MK-ULTRA records from the DNSA collection and potential congressional oversight hearings in response to The Intercept's reporting.

Allied Intelligence

Estonias Foreign Intelligence Chief Says Russia Cannot Replenish Fallen Soldiers

In an interview published April 26 by EUobserver, Kaupo Rosin, Director General of the Estonian Foreign Intelligence Service since 2022, stated Russia faces a serious battlefield problem in replenishing troops. He assessed there is currently no reason to panic regarding a Russian attack on NATO in 2026, provided collective defense remains robust, as Russia's military resources are consumed by the Ukraine invasion. Rosin characterized most Russian soldiers as having low IQ levels and being convicts from the lowest strata of society. Estonia currently allocates over 5 percent of GDP to defense and advocates this as a minimum for European nations.

Analyst Note: Rosin's public judgment carries unusual weight for a single-service assessment. Estonian Foreign Intelligence Service (FIS) operates closer to the Russian military intelligence target than most allied services. The characterization of losses outpacing recruitment, though undermined by his editorializing about Russian soldiers' backgrounds, is the most explicit force-generation failure claim on record from a serving European intelligence chief. His 2026 no-attack reassurance, conditioned on collective defense holding, simultaneously manages Baltic public anxiety and applies pressure to NATO members spending below Estonia's 5-percent threshold. The public reassurance may instead reflect a deliberate effort to counter domestic unease rather than a dispassionate intelligence judgment, masking classified indicators of faster Russian reconstitution. Corroboration from Institute for the Study of War (ISW) or UK MoD that losses have exceeded recruitment for three consecutive months would elevate this to allied consensus, and such corroboration is likely within the coming weeks.

Sources:

Watch Items - Estonian FIS assessment of Russian manpower shortages: watch for ISW or UK MoD corroboration that Russian losses have exceeded recruitment for three consecutive months. - Germany may restrict Signal desktop on parliamentary systems following the phishing compromise, with implications for secure legislative communications across NATO.

COLLECTION GAPS

UNCLASSIFIED // OPEN SOURCE