KEVSEC Weekly Intelligence Report — August 17, 2026

This KEVSEC Weekly Intelligence Report covers the period ending August 17, 2026. Active threat sensor infrastructure recorded 11 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. Full analysis across cyber threat intelligence, infrastructure defense metrics, vulnerability landscape, global events, and government intelligence follows.

Over the past seven days, the KEVSEC honeypot and network sensor infrastructure logged 11 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.

At the autonomous system level, the highest-volume sources were: Google Cloud Platform (110 hits); DigitalOcean LLC (20 hits); TECHOFF SRV LIMITED (14 hits); Microsoft Corporation (12 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-2026-33267 (CVSS 10.0) — Improper Input Validation vulnerability in Apache Traffic Server.

This issue affects Apache Traffic Server: from 9.2.0 through 9.2.14, from. Defenders are advised to prioritize patching based on EPSS exploitation probability scores alongside raw CVSS severity ratings.

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; 228 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.

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-2026-33267 (CVSS 10.0, CRITICAL, published 2026-07-29): Improper Input Validation vulnerability in Apache Traffic Server.

This issue affects Apache Traffic Server: from 9.2.0 through 9.2.14, from 10.1.0 through 10.1.3.

Users are recommended to upgrade to EPSS exploitation probability: 34.0%.

CVE-2026-55971 (CVSS 9.8, CRITICAL, published 2026-07-27): Heap-based Buffer Overflow vulnerability in Apache Thrift C++ bindings.

This issue affects Apache Thrift: before 0.24.0.

Users are recommended to upgrade to version 0.24.0, which fixes the issue. EPSS exploitation probability: 104.0%.

CVE-2026-48144 (CVSS 9.1, CRITICAL, published 2026-07-27): Improper Validation of Certificate with Host Mismatch vulnerability in Apache Thrift c_glib bindings.

This issue affects Apache Thrift: before 0.24.0.

Users are recommended to upgrade to version 0.2 EPSS exploitation probability: 42.0%.

CVE-2026-58023 (CVSS 9.1, CRITICAL, published 2026-07-27): Out-of-bounds Read vulnerability in Apache Thrift c_glib bindings.

This issue affects Apache Thrift: before 0.24.0.

Users are recommended to upgrade to version 0.24.0, which fixes the issue. EPSS exploitation probability: 108.0%.

CVE-2026-58662 (CVSS 9.1, CRITICAL, published 2026-07-27): Improper Validation of Specified Quantity in Input, Out-of-bounds Read vulnerability in Apache Thrift C++ bindings.

This issue affects Apache Thrift: before 0.24.0.

Users are recommended to upgrade EPSS exploitation probability: 115.0%.

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.

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 earthquake (Magnitude 5.7M, Depth:48.947km) in Indonesia 17/08/2026 00:46 UTC, 840 thousand in MMI IV. (GREEN alert, Indonesia): EQ event logged Mon, 17 Au.

⚠ Green forest fire notification in Kazakhstan (GREEN alert, Kazakhstan): WF event logged Mon, 17 Au.

⚠ Green earthquake (Magnitude 5.5M, Depth:10km) in Indonesia 16/08/2026 06:51 UTC, 480 thousand in MMI IV. (GREEN alert, Indonesia): EQ event logged Sun, 16 Au.

⚠ Green earthquake (Magnitude 6.1M, Depth:188.148km) in Vanuatu 16/08/2026 03:20 UTC, 5 thousand in MMI IV. (GREEN alert, Vanuatu): EQ event logged Sun, 16 Au.

USGS seismic monitoring recorded 5 significant earthquakes this week. The highest-magnitude event was a M7.7 near 68 km NNW of Ende, Indonesia, registered 2026-08-14 4:58 PM 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 16 active wildfires across the United States covering approximately 2,658,601 total acres. Largest active incidents: BIG GRASS (OR, 578,448 acres, 72% contained); 0445 CROSSWHITE (OR, 344,916 acres, 87% contained); Coleman Creek (OR, 308,721 acres, 90% contained); Tartar (ID, 158,027 acres, 98% 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=1.33 (1P) — conditions are quiet (nominal conditions). 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.

White House and presidential activity tracked during this reporting period reflects the following developments: "Expanding Capabilities to Combat Transnational Cyber-Enabled Crime" | "National Substance Use Primary Prevention Month, 2026" | "Trump sued over service selling faster access to Truth Social posts - NPR". 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. "11 Defendants Charged in Dismantling of Decade-Long Nationwide Marriage Fraud Scheme - Department of Justice (.gov)" (DOJ, Wed, 12 Au); "Six Medicaid Providers Facing Fraud, Theft Charges - Ohio Attorney General (.gov)" (DOJ, Tue, 23 Ju). Law enforcement actions, indictments, and agency press releases are monitored as indicators of domestic threat mitigation and enforcement priority shifts.

Additional government intelligence from Pentagon captures the broader policy and operations picture. "House narrowly passes $1.15 trillion defense bill, approving raises for troops, adopting Department of War name - stripes.com"; "House Republicans pass hard-right defense bill, handing wins to Trump - Politico"; "House passes $1.15 trillion defense bill in near party-line vote - Military Times".

The 119th Congress is currently in recess. Current composition: House 220R/213D; Senate 53R/45D. Active legislative items tracked this week: "A Growing Share of Federal Spending Escapes Regular Congressional Review - Bipartisan Policy Center"; "House passes temporary government funding bill months early - PBS"; "House passes Pentagon funding bill and a blueprint to unlock new dollars for Iran war - NPR"

Automated red team assessment of https://kevsec.com completed 2026-08-16T03:00:07.916903. 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.

From an analyst standpoint, this reporting period reinforces several persistent trends worth monitoring. Threat actor activity maintained baseline levels without notable escalation.

Vulnerability priority should center on CVE-2026-33267 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: 16 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.