Improper validation of unsafe equivalence in input could allow an attacker with high privileges to reclaim the ring buffer queue status page, potentially resulting in loss of guest memory integrity.
CVE-2026-120979 is a high-severity vulnerability in Arista Networks EOS affecting customers using Arista Pathfinder with BGP-Link State configured for Dynamic Path Selection. An authenticated attacker with a BGP-LS session can send crafted packets to cause incorrect WAN path selection, disrupting traffic forwarding. The issue is due to improper validation of unsafe equivalence in input. Fixed versions include 4.32.0F and 4.33.0F.
CVE-2026-116870 is a vulnerability in Octopus Deploy's Octopus Server where OIDC token issuer validation was inconsistently performed. This flaw could allow a token from a non-legitimate issuer to be accepted as if it belonged to a legitimate trusted service account. Exploitation requires a service account with at least one OIDC identity configured and some knowledge or guess of that identity's details. The issue affects multiple versions of Octopus Server prior to fixed releases in 2026.1.11781, 2026.2.13438, and 2026.3.15816. The vulnerability has a high severity CVSS score of 8.7. No known exploits in the wild have been reported.
A vulnerability in webpack-dev-server versions 5.1.0 up to but not including 5.2.7, and 6.0.0 up to but not including 6.1.0, allows unauthenticated attackers to cause the server to disclose sensitive information via error messages. This occurs because the open-editor endpoint improperly validates the Host header, leading to an exception that reveals absolute filesystem paths, module locations, and source line numbers in HTTP 500 responses. The server continues running without denial of service. The issue is fixed in versions 5.2.7 and 6.1.0.
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.56.0 until 1.107.6 and 2.44.0, applications that opt attacker-influenced URLs into local network access through FileUrl with force_download='allow-local' or web_fetch_tool with allow_local_urls=True can bypass the cloud-metadata blocklist by appending an IPv6 zone identifier to an IPv6 metadata address. IPv6Address equality and hashing include the zone identifier, so the blocklist comparison fails even though the network stack ignores the zone on a non-link-local destination and reaches the metadata service, potentially exposing cloud IAM credentials. The opt-in settings are disabled by default, and the issue requires an IPv6-enabled environment. This issue is fixed in versions 1.107.6 and 2.44.0.
Pydantic AI is a Python agent framework for building applications and workflows with Generative AI. From 1.77.0 until 1.107.6 and 2.44.0, the local web_fetch_tool and the WebFetch local fallback compare blocked_domains entries with a URL hostname before both values are normalized to the form used by getaddrinfo. An attacker-influenced model can use an equivalent IDNA spelling, non-ASCII label separator, case variation, or trailing root label that resolves to a blocked host but does not match the configured string, causing the application to fetch that host with its own privileges. allowed_domains fails closed for unmatched spellings, and private-IP and cloud-metadata protections remain effective. This issue is fixed in versions 1.107.6 and 2.44.0.
In affected versions of Octopus Server, an authenticated user with permission to modify roles could bypass the protections preventing access abuse resulting in privilege escalation. It was possible for the built-in role to be weakened and the attacker's account added to a privileged team. This was achievable due to improper validation of unsafe equivalence in inputs.
Handlebars provides the power necessary to let users build semantic templates. From 3.0.0 until 4.7.10, Handlebars lookupProperty returns Function.prototype.constructor before applying the prototype-access deny list because constructor is an own property of Function.prototype. When an attacker can render a controlled template with allowProtoMethodsByDefault enabled and an accessible function in the template context, the template can traverse from that function through its prototype to Function.prototype and then obtain the Function constructor through the own-property bypass. This permits attacker-controlled JavaScript to execute with the server application's privileges. This issue is fixed in version 4.7.10.
Zilliz Attu versions from 2.6.5 up to but not including 3.0.0 contain a Server-Side Request Forgery (SSRF) vulnerability in the Playground feature. This vulnerability arises from improper validation of unsafe equivalence in input, allowing proxying of requests to private IP addresses. The CVSS 3.1 base score is 4.0, indicating medium severity. There is no indication of active exploitation in the wild. No official patch or remediation details are provided in the available data.
A vulnerability in JetBrains TeamCity before versions 2026.2, 2026.1.4, and 2025.11.8 allows an attacker to take over administrator accounts via the password reset functionality. This issue is tracked as CVE-2026-100255 and is associated with CWE-1289. The vulnerability has a high severity rating with a CVSS 3.1 score of 8.1.
Improper Validation of Unsafe Equivalence in Input vulnerability in Liquid Web / StellarWP GiveWP allows Authentication Bypass.
This issue affects GiveWP: from n/a through 4.16.9.
Contrast (Edgeless Systems) through 1.20.0 performs unanchored suffix matching when selecting per-registry configuration in the imagepuller. Config.registryFor strips a single trailing dot and then uses strings.HasSuffix(hostname, fqdn) without requiring a DNS label boundary, so a registry entry such as [registries."ghcr.io."] is also applied to any host whose name merely ends in that byte sequence, including attacker-registered domains such as evilghcr.io. When an image or layer is pulled from such a sibling domain, the imagepuller sends the configured Authorization header (basic auth, registry token, or identity token), trusts the configured custom CA bundle, follows the configured mirror, and honours insecure-skip-verify (disabling TLS verification) for that host. Image integrity is not affected, as image bytes remain pinned by digest in the policy and are validated after the pull. Configurations that use a leading dot (e.g., [registries.".example.registry"]) are unaffected.
Improper validation of pod identifier uniqueness in aws-network-policy-agent in Amazon EKS Network Policy Agent before v1.4.0 might allow an authenticated remote user to bypass NetworkPolicy enforcement on co-located pods in other namespaces via crafted pod and namespace names that produce pod identifier collisions.
To remediate this issue, users should upgrade to Amazon EKS Network Policy Agent 1.4.0 or later and Amazon VPC CNI Managed Add-on v1.22.4 or later (which includes Network Policy Agent v1.4.0).
Improper Validation of Unsafe Equivalence in Input in ZenHive mpp allows an unauthenticated remote client to pass the Tempo duplicate-submission gate twice with one signed transaction.
MPP.Methods.Tempo reserves the pre-broadcast dedup slot on the caller-supplied hex in reserve_hash_atomic/2, keyed through store_key/1 on tx.raw rather than on a canonical form of the transaction. The deserializer stores the caller's hex verbatim and accepts both recovery-id encodings, so one signed transaction submitted once with v=27 and once with v=0 yields two distinct reserve keys, and both pass the reserve and reach the broadcast path. The plug-level credential replay store is deliberately carved out for tempo in lib/mpp/replay.ex, leaving this reserve as the only gate, and the post-broadcast mark writes the canonical hash key that the raw-keyed reserve never reads.
What the duplicate submission yields depends on the node: a nonce-reuse rejection fails closed, while a node that answers with the canonical hash for an already-known transaction returns a second valid Payment-Receipt for a single on-chain payment.
This issue affects mpp: from 0.2.0 before 0.16.2.
CVE-2026-89049 is a critical server-side request forgery (SSRF) vulnerability in Amazon AWS Systems Manager Agent (SSM Agent) versions before 3.3.4851.0. It arises from improper validation of equivalent address representations in the port forwarding functionality, allowing an authenticated remote user to bypass the remote destination denylist and access link-local endpoints. This can lead to unauthorized retrieval of temporary IAM role credentials from a managed instance, enabling actions with those permissions from outside the instance. The issue is fixed in version 3.3.4851.0 and later.
CVE-2026-76977 is a medium severity vulnerability in SAP UI5's Frame Options Allowlist. The issue arises from insufficient validation of the parent frame's origin against the configured allowlist, allowing an unauthenticated attacker to host a malicious page that can bypass framing restrictions. If an authenticated user visits the attacker's page and interacts with it, the attacker could trick the user into performing unintended actions. The impact is limited to integrity with no confidentiality or availability impact.
CVE-2026-74994 is an incorrect authorization vulnerability in the mod_auth module of Erlang OTP's inets httpd server. When configured with dets or mnesia authentication backends and multiple directory configuration blocks, all directory blocks share a single user/group namespace. This causes a user authorized for one protected directory to be accepted for all other protected directories on the same server instance. The vulnerability affects multiple OTP and inets versions prior to specific fixed releases.
URI versions before 5.36 for Perl encode non-NFC host names to non-standard punycode labels via missing normalization in nameprep.
nameprep lowercases each host label but performs no Unicode normalization. IDNA requires a label to be normalized to Form C before it is encoded (RFC 5891), so a label that is not already in NFC is encoded to a different A-label than its normalized form. A label built from the precomposed Devanagari sequence U+0958 U+093E encodes to xn--72b5c without normalization but to xn--11b2fg after NFC normalization, and xn--72b5c does not round-trip back to the original label.
Any caller that reads host() from a URI built from untrusted input and uses it for a security decision (an allow or deny list, an SSRF filter, deduplication, a cache key) sees the non-standard label, while a client that fetches the same URL resolves the NFC form, so the check and the fetch can disagree about the host.
Date::Manip versions through 7.00 for Perl return corrupted dates via non-ASCII decimal digits that pass the numeric range tests in check.
The parse regexes capture year, month and day with the `\d` shorthand, which on a character string matches the whole Unicode decimal digit property `\p{Nd}` and not just `[0-9]`. Date::Manip::Base::check then validates the captured fields with numeric comparisons alone (`$y<1 || $y>9999`, `$m<1 || $m>12`, `$d<1 || $d>$days`), and _parse_check stores the numified fields (`$y+0`). Perl truncates a string at the first character that is not an ASCII digit, so a field whose leading characters are ASCII digits numifies to an in-range prefix and satisfies every test: a year field of three ASCII digits followed by U+0664 ARABIC-INDIC DIGIT FOUR numifies to 202, giving the year 0202, and one non-ASCII digit in the month or day field shifts those fields the same way. The hour, minute and second fields match explicit ASCII character classes (`0?[0-9]`, `[0-5][0-9]`) and do not shift, though a non-ASCII digit in a fractional hour or minute field truncates the fraction.
Any caller that passes an untrusted character string to ParseDate() or Date::Manip::Date->parse() can get back a date that differs from the string it parsed, with no parse error. Where the parsed date gates logic such as an expiry check or a retention window, the shift goes unnoticed.