KEVSEC // INTELLIGENCE REPORT // OPEN SOURCE // UNCLASSIFIED
Weekly Intelligence Briefing · Week 40, 2026 · KEVSEC Operations Center

KEVSEC Weekly Intelligence Report — October 5, 2026

October 5, 2026 | 1860 words | 8 sections
threat-intelligence cves government global-events infrastructure state-actors vulnerabilities red-team wildfires
← All Reports
// Section 1 of 8

Executive Summary

This KEVSEC Weekly Intelligence Report covers the period ending October 5, 2026. Active threat sensor infrastructure recorded 12 inbound probe attempts from 3 unique source IPs across the reporting window. The SSH tarpit neutralized 429 unique scanner IPs, consuming 21.7 hours of adversary connection time. 13 IPs were permanently banned following honeypot engagement. The cumulative nftables blocklist now stands at 18,820,417 permanently blocked addresses.

Key intelligence findings this cycle: 20 vulnerability disclosures tracked by NVD including 5 Critical-rated CVEs; state-affiliated probe activity attributed to Romania, China infrastructure. Full analysis across cyber threat intelligence, infrastructure defense metrics, vulnerability landscape, global events, and government intelligence follows.


// Section 2 of 8

Cyber Threat Intelligence

Over the past seven days, the KEVSEC honeypot and network sensor infrastructure logged 12 inbound probe attempts from 3 unique IP addresses — a steady tempo consistent with broad opportunistic reconnaissance campaigns. The activity spans automated SSH brute-force sweeps, HTTP vulnerability scanners, and targeted probes against exposed management interfaces, reflecting the persistent baseline noise of internet-facing infrastructure.

State-affiliated infrastructure contributed a measurable share of inbound activity. The top nation-state sources by probe volume were: 🇷🇴 Romania (18 recorded hits via M247 Europe SRL); 🇨🇳 China (3 recorded hits via Beijing Volcano Engine Technology Co., Ltd.). Attribution to state actors is based on autonomous system ownership as resolved through BGP routing data — it reflects infrastructure origin, not necessarily direct government sponsorship of individual scanning operations.

At the autonomous system level, the highest-volume sources were: Google Cloud Platform (343 hits); Omegatech LTD (48 hits); TC DATACENTER LIMITED (24 hits); Miteflux Technologies Ltd (21 hits). The presence of cloud hosting providers and residential ISPs within the top sources underscores how threat actors leverage commodity infrastructure to distribute scanning operations and complicate attribution.

The vulnerability front saw 6 notable disclosures this reporting period, including 5 rated Critical and 1 rated High. The highest-priority finding was CVE-2026-65381 (CVSS 10.0) — A validation issue existed in the entitlement verification. This issue was addressed with improved validation of the process entitlement. Th. Defenders are advised to prioritize patching based on EPSS exploitation probability scores alongside raw CVSS severity ratings.


// Section 3 of 8

Infrastructure Defense Metrics

The KEVSEC endlessh SSH tarpit has processed 429 unique adversary IP addresses and consumed a cumulative 21.7 hours of attacker connection time since deployment — including 21.7 hours in this reporting period alone. The tarpit operates by accepting SSH connections on port 22 and deliberately delaying the protocol handshake, tying up scanner threads and consuming attacker resources without providing any authentication surface. Tarpit data feeds directly into the nftables permanent blocklist — identified scanner IPs are automatically flagged for injection into the deny ruleset, preventing future connection attempts.

The HTTP honeypot layer logged 47 probe attempts from 4 unique IPs during this period. 13 IPs triggered automatic permanent bans via the honeypot engagement pipeline. Monitored honeypot paths include common WordPress admin panels, phpMyAdmin instances, environment file endpoints, Git configuration directories, and known CVE-targeted paths — all designed to catch automated exploitation frameworks and manual attackers alike.

Fail2ban enforcement triggered across 4 active jails this cycle. Top jail activity by ban count: sshd-root (10 bans); honeypot-probe (7 bans); honeypot-trap (7 bans); sshd (6 bans). The sshd and sshd-root jails represent brute-force credential attacks against the management interface; honeypot-probe and honeypot-trap jails capture HTTP-layer exploitation attempts.

The nftables blacklist enforced 4 logged block events across 1 enforcement sources. Block sources: HONEYPOT (4).

Cumulative platform metrics: 18,820,417 IPs permanently blocked in the nftables deny ruleset; 274 total probe events logged across all sensor layers since deployment. The permanent blocklist is maintained as a compiled nftables set and is reloaded hourly from the consolidated blacklist — ensuring minimal performance overhead while providing comprehensive pre-connection blocking.


// Section 4 of 8

Vulnerability Landscape

The National Vulnerability Database published or updated 8 significant vulnerabilities during this reporting window — 5 rated Critical and 3 rated High. EPSS (Exploit Prediction Scoring System) probability data, where available, is included to help prioritize remediation efforts against actively exploited or exploit-imminent weaknesses.

CVE-2026-65381 (CVSS 10.0, CRITICAL, published 2026-09-14): A validation issue existed in the entitlement verification. This issue was addressed with improved validation of the process entitlement. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15. EPSS exploitation probability: 35.0%.

CVE-2026-81642 (CVSS 9.8, CRITICAL, published 2026-09-16): In NLnet Labs Unbound up to and including 1.26.0, a vulnerability was found in the DNSSEC validator that enables denial of service and possible remote code execution as a result of digesting DNSKEYs. EPSS exploitation probability: 96.0%.

CVE-2026-82717 (CVSS 9.8, CRITICAL, published 2026-09-16): In NLnet Labs Unbound up to and including 1.26.0, a vulnerability was found in that can progressively corrupt heap memory and under certain systems and compilation options could lead to remote code ex EPSS exploitation probability: 78.0%.

CVE-2026-43790 (CVSS 9.1, CRITICAL, published 2026-09-14): The issue was addressed with improved memory handling. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. A remote attacker may be able to cause unexpected system termi EPSS exploitation probability: 72.0%.

CVE-2026-18119 (CVSS 9.0, CRITICAL, published 2026-09-14): Concrete CMS below 9.5.3 did not sanitize custom style values in the Block Design dialog before writing them into page CSS via a DOM sink, permitting stored cross-site scripting. An editor-level user EPSS exploitation probability: 31.0%.

Organizations should review their asset inventory against the affected vendors and platforms listed above. Where EPSS scores exceed 10%, treat remediation as urgent regardless of CVSS rating, as probability of observed in-the-wild exploitation is elevated.


// Section 5 of 8

Global Security Events

The Global Disaster Alerting and Coordination System (GDACS) flagged 6 significant events during this reporting period, 0 carrying red-level alerts and 0 at the orange threshold. Red alerts indicate high humanitarian impact potential requiring immediate international response coordination.

⚠ Green flood alert in Thailand (GREEN alert, Thailand): FL event logged Mon, 05 Oc.

⚠ Green flood alert in Mexico (GREEN alert, Mexico): FL event logged Sat, 26 Se.

⚠ Green notification for tropical cyclone KOGUMA-26. Population affected by Category 1 (120 km/h) wind speeds or higher is 0 (0 in Tropical Storm). (GREEN alert, ): TC event logged Mon, 05 Oc.

⚠ Green earthquake (Magnitude 5.5M, Depth:10km) in Volcano Islands, Japan Region 04/10/2026 11:12 UTC, [unknown]. (GREEN alert, Volcano Islands, Japan Region): EQ event logged Sun, 04 Oc.

USGS seismic monitoring recorded 2 significant earthquakes this week. The highest-magnitude event was a M4.7 near 5 km NNE of Korumburra, Australia, registered 2026-10-02 11:01 AM CT. Events above M6.0 in densely populated regions are tracked for potential secondary effects on critical infrastructure, supply chains, and regional stability.

NIFC wildfire monitoring reports 13 active wildfires across the United States covering approximately 983,879 total acres. Largest active incidents: LITTLE GIANT (WA, 172,902 acres, 93% contained); SINLAHEKIN (WA, 166,007 acres, 99% contained); Second Flat (OR, 105,854 acres, 99% contained); Aspen Acres (CO, 102,004 acres, 82% contained). Active wildfires in populated regions are tracked for potential air quality impacts, infrastructure disruption, and emergency services strain that may degrade local response capabilities.

NOAA Space Weather Prediction Center reports current geomagnetic activity at Kp=3.0 (3Z) — conditions are active (G1-G2 — minor power grid fluctuations, possible HF impacts). No active SWPC alerts during this reporting period. Elevated geomagnetic activity (Kp≥5) can impact HF communications, satellite navigation accuracy, and high-latitude power grid stability — relevant for emergency communications planning and critical infrastructure operators.


// Section 6 of 8

Government & Political Intelligence

White House and presidential activity tracked during this reporting period reflects the following developments: "Here’s The White House’s 308-Word AI ‘Accord’ In Full—Including Misspelling - Forbes" | "Trump’s funding cancellation violates budget law, GAO says - Roll Call" | "Appropriators seek ways to curb Trump’s spending cancellations - Roll Call". Presidential movements and official actions are catalogued from White House RSS feeds and corroborating news sources to provide an unfiltered activity record.

Department of Justice and FBI reporting highlights 2 notable developments this week. "Department of Justice Indicts Ten Aliens for Voter Fraud in Minnesota - Department of Justice (.gov)" (DOJ, Wed, 30 Se); "Former SPLC official charged in Justice Department's broader criminal case against group - PBS" (DOJ, Wed, 12 Au). Law enforcement actions, indictments, and agency press releases are monitored as indicators of domestic threat mitigation and enforcement priority shifts.

Additional government intelligence from State Dept, Pentagon, CENTCOM captures the broader policy and operations picture. "Operation Inherent Resolve is Over After 12 Years: US Withdraws Forces From Iraq - Military.com"; "House votes to adopt Department of War renaming as part of annual defense bill - PBS"; "Secretary Rubio’s Travel to Los Angeles - U.S. Department of State (.gov)".

The 119th Congress is currently in session. Current composition: House 220R/213D; Senate 53R/45D. Active legislative items tracked this week: "Republicans Advance Largest Pentagon Budget in History while Cutting Billions from Domestic Investments - democrats-appropriations.house.gov"; "House passes Russia sanctions, sending bill named for Lindsey Graham to Trump's desk - NPR"; "Fiscal 2026 funding discussions to dominate House, Senate agenda - Roll Call"


// Section 7 of 8

Red Team Assessment

Automated red team assessment of https://kevsec.com completed 2026-10-04T03:00:09.042246. Scan returned 0 Critical, 0 High, 3 Medium, 1 Low, and 40 informational findings. Overall posture: CLEAN. The KEVSEC platform performs weekly automated security assessments against its own externally-facing infrastructure, correlating findings against known-good baseline configurations to identify drift, misconfigurations, or newly exposed attack surface.

Actionable findings this cycle:

[MEDIUM] Auth — Login rate limiting may not be active (test inconclusive from localhost): Could not trigger lockout in 7 attempts — verify _check_rate_limit is wired to login route

[MEDIUM] SSH — Password authentication enabled on SSH: PasswordAuthentication yes

[MEDIUM] SSH — Password authentication enabled on SSH: PasswordAuthentication yes

24 security controls confirmed in-place during this assessment cycle. Verified controls include: strict-transport-security; content-security-policy; x-frame-options; x-content-type-options; referrer-policy; permissions-policy; Env file exposed: /.env → HONEYPOT ACTIVE; Git repo exposed: /.git/config → HONEYPOT ACTIVE.

Red team findings are triaged weekly. Critical and High findings trigger immediate remediation workflows. Medium findings are tracked in the configuration backlog. All findings are re-assessed on the following cycle to confirm remediation.


// Section 8 of 8

Analyst Notes

From an analyst standpoint, this reporting period reinforces several persistent trends worth monitoring. Scanning pressure from Romania, China-attributed infrastructure remains consistent — no significant surge, but no decline either.

Vulnerability priority should center on CVE-2026-65381 and any assets running affected software versions. Organizations running unpatched internet-facing systems against this week's Critical disclosures should treat those systems as compromised until verified otherwise.

The political intelligence picture continues to evolve with frequent executive activity. Downstream policy effects on regulatory posture, federal IT procurement, and agency operational tempo remain areas to watch in subsequent reporting cycles.

Infrastructure defense posture remains active. The tarpit consumed 21.7 hours of adversary time this week alone — a passive but meaningful drain on automated scanning resources targeting this platform. Consistent tarpit engagement without escalation suggests adversaries have not flagged this infrastructure as a high-value target warranting manual follow-up, though this assessment should be revisited if targeted probe patterns emerge.

Environmental conditions bear monitoring: 13 active wildfires are currently burning across the US. Wildfire incidents of sufficient scale can stress emergency communications infrastructure, trigger emergency management coordination activity, and create information operations opportunities around disaster response narratives.

KEVSEC reports are generated weekly from live sensor data, open-source intelligence feeds, and public government sources. Data reflects conditions at the time of generation and should be supplemented with direct-source verification for operational decisions.