Security1 distinct publisher2 min readPublished
Two flaws in Applied Systems Engineering's ASE2000 V2 test set, one of them an eight-year-old log4net XXE, let an attacker read local files and sit inside the TLS session engineers use to check SCADA traffic. Version 2.38 fixes both.
The Watch · Security desk

Compiled by The WatchSomething wrong?How this is made
The XXE half carries a precondition the advisory states plainly: log4net has to parse a configuration file the attacker controls [3]. That is a promotion of access someone already has, not a way in. Whoever can write into the ASE2000 installation directory converts that write into arbitrary local file reads and outbound requests from the host [4], which is why the first interim measure is a permissions change on the install directory and its config files rather than anything on the network [8].
CVE-2026-18717 is the one that works from the wire. In versions 2.35 through 2.37 the IEC 60870-5-104 TLS client does not validate certificate error conditions correctly, so an attacker positioned to intercept can present a certificate, complete the handshake as the trusted peer, and read or modify what crosses the session [5][6]. A test set that accepts a forged peer is describing the attacker's traffic to the engineer reading its output, and on a commissioning link there is usually no second source to compare against.
The two bugs do not cover the same ground. Counting the listed range, 2.25 through 2.37 is 13 releases [10]; the certificate flaw is scoped to three of them [11], leaving ten builds exposed only to the log4net issue [12]. Both close on the same version, so the split does not buy anyone a smaller upgrade: every customer between 2.25 and 2.37 is told to move to 2.38 or later [7].
Upstream log4net fixed the entity handling in 2.0.10 [3]. Version 2.38 ships 3.3.1.0, well past the minimum that would have closed it [14]. The identifier attached to that flaw is CVE-2018-1285, eight years older than the certificate CVE published alongside it [13]. That gap is the interesting number here: the fix had been available upstream the whole time, and the clock was running on the vendor rebuild.
The advisory as supplied lists no CVSS metrics and no report of exploitation in the wild [9]. So this sits in the inventory column. The questions worth answering are which hosts run 2.35 or later, who has write access to their program directories, and whether the IEC 104 TLS sessions those hosts open cross a network the operator does not own [8].
Ranked by verification strength, evidence, and original report placement.
CISA published an advisory on the Applied Systems Engineering ASE2000 V2 Communications Test Set covering CVE-2018-1285 and CVE-2026-18717, with affected versions listed as ASE2000 >=2.25 through <=2.37.
CISA lists the ASE2000 V2 Communications Test Set as deployed worldwide across the Chemical, Critical Manufacturing, Energy, and Water and Wastewater critical infrastructure sectors, with the company headquartered in the United States.
CVE-2018-1285: Apache log4net versions before 2.0.10 do not disable XML external entities when parsing log4net configuration files, allowing XXE based attacks in applications that accept attacker controlled log4net configuration files. Relevant CWE is CWE-611.
CISA states successful exploitation could allow an attacker to read or write arbitrary local files, cause the application to issue outbound network requests, or intercept the connection to impersonate the trusted peer, complete the TLS handshake, and read or modify the protected communications.
ASE2000 2.35 through 2.37 is vulnerable to an improper certificate validation vulnerability (CVE-2026-18717) which may allow an attacker to impersonate the trusted peer, complete the TLS handshake, and read or modify protected communications.
ASE/Kalkitech provides version 2.38, which fixes both vulnerabilities: the bundled log4net library is upgraded to 3.3.1.0 and the IEC 60870-5-104 TLS client certificate validation logic is corrected to ensure proper validation of certificate error conditions.
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 27, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
security
Siemens IoT2050 gateways ship a Node-RED interface that asks nobody for a password1 distinct publisher
security
CISA's Ebyte advisory carries no fixed version, because the vendor stopped answering1 distinct publisher
security
Johnson Controls console holds passwords in cleartext memory, and the fix line names two versions1 distinct publisher
security
CISA's water-sector answer is an inventory: 100-plus exposed systems, most of them PLCs1 distinct publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Single authoritative advisory, no severity metrics
Every factual element is traceable to a first-party CISA ICS advisory carrying assigned CVE identifiers, CWE mappings, explicit version ranges, a named reporter, and vendor-supplied remediation detail - strong provenance for the existence and nature of both flaws. It is capped by being one source with no CVSS metrics under either Metrics heading, no technical proof-of-concept, no independent corroboration, and an unexplained mismatch between the product-level and per-vulnerability version ranges.
Fix shipped; real-world uptake unmeasured
There is concrete supply-side adoption evidence - a fixed release (2.38) exists, is available from the vendor, and upgrading is stated as required for every install in 2.25 through 2.37 - plus a qualitative statement that the product is deployed worldwide across four critical infrastructure sectors. What is absent is any demand-side measurement: no install-base numbers, no count of affected customers, no upgrade uptake, and no telemetry or scan data on exposed hosts, so the score reflects documented remediation availability rather than demonstrated remediation.
Slightly understated by the absent severity metrics
The advisory's language is restrained and matches its evidence: impacts are stated conditionally ('could allow', 'may allow'), no exploitation-in-the-wild is asserted, and the cluster headline and dek stay within the advisory's own impact chain of arbitrary local file read/write and TLS peer impersonation. The mild negative reflects understatement rather than inflation - a full TLS-interception flaw in an IEC 60870-5-104 client used on critical infrastructure networks is published with empty Metrics sections and no severity score, and the persistence of an eight-year-old log4net XXE across 13 releases is left without comment.
Government disclosure carrying vendor-drafted remediation
The publisher is a government agency with a coordinated-disclosure mandate and no commercial stake, and the finding is credited to an outside reporter, which lowers promotional incentive. Offsetting that, the remediation and mitigation text reads as vendor-supplied ASE/Kalkitech language that routes readers to the vendor's site and support address, frames the flaws as fully resolved by a version upgrade, and benefits from omitting severity scoring, historical exposure duration, and any account of the range mismatch.
High confidence in the facts, low in the risk sizing
Confidence is high that both vulnerabilities exist as described, that the affected ranges and CWE mappings are as stated, and that version 2.38 is the vendor's fix - these come directly from the issuing authority with CVE identifiers attached. It is held down because the cluster has a single publisher, no severity metrics, no exploitation data, no install-base figures, and an internal range inconsistency, so any judgement about how urgent or widespread the exposure actually is rests on inference rather than the supplied record.