Information Disclosure

Out-of-bounds read in the Linux kernel's PPP subsystem via the ppp_handle_ipcp function allows a remote PPP peer to supply a crafted IPCP packet with an attacker-controlled length field, triggering an OOB read in kernel memory. Systems running any kernel version with PPP support enabled and an active PPP link are potentially affected. Reported by Ubuntu; no patch version, CVSS score, or KEV status is confirmed at time of analysis.

RSA-PSS signature verification in a TLS library's CertificateVerify processing validates only the mandatory 0xbc trailer byte rather than performing complete cryptographic verification of the PSS-encoded signature structure. This flaw, reported by Ubuntu, could allow a remote attacker to forge RSA-PSS CertificateVerify messages during a TLS handshake, potentially enabling authentication bypass without possession of a valid private key. No public exploit or CISA KEV entry is associated with this CVE at time of analysis.

Unbounded recursion in Mongoose's DER parsing function mg_der_find_oid allows a remote attacker to trigger a stack overflow by sending crafted ASN.1/DER data containing deeply nested constructed tags, likely resulting in a denial of service through process crash. Affected deployments include any application using Mongoose's TLS or certificate-parsing code paths that process externally-supplied DER-encoded input. No public exploit has been identified at time of analysis, and this vulnerability was reported via the Ubuntu vendor channel.

Off-by-one out-of-bounds read in an HTTP chunked transfer encoding parser - specifically in the CRLF check within a skip_chunk code path - affects software distributed via Ubuntu. The flaw allows network-accessible parsing code to read one byte past a valid buffer boundary when processing chunked HTTP request or response data. Depending on memory layout, this may yield information disclosure or a process crash; no public exploit has been identified at time of analysis and no CVSS vector is available to further qualify scope.

CVE-2026-62289 is an Ubuntu security update with no public technical details available at time of analysis. The sole intelligence source is a vendor report from Ubuntu, with no description, CVSS score, CWE classification, or affected component provided. The nature of the vulnerability, the affected subsystem, and attacker impact cannot be determined from available data.

Unbounded memory growth in the FRN module (GHSA-qccg-7pw6-787v) can lead to resource exhaustion in affected deployments. The vulnerability was reported via Ubuntu's vendor channel and is associated with a GitHub Security Advisory, suggesting it is a library or kernel module-level issue. No CVSS score, CWE classification, or version details have been published at time of analysis, making precise impact assessment limited.

Out-of-bounds read in the IPv6 extension header receive path (`rx_ip6`) allows a remote unauthenticated attacker to trigger kernel memory disclosure or a denial-of-service condition via a crafted IPv6 packet with a malicious 16-bit extension header length field that wraps around during bounds calculation. Reported by Ubuntu vendor channels, this is inferred to affect the Linux kernel's IPv6 network stack, though the exact upstream component is not confirmed from available data. No public exploit code has been identified at time of analysis, and KEV status is not confirmed.

Ubuntu's IPv6 network stack fails to validate the MTU value carried in NDP Router Advertisement (RA) messages, allowing any arbitrary MTU value to be accepted by the host. An attacker with the ability to send crafted IPv6 RA packets on the local network segment can set an excessively low or malformed MTU, disrupting IPv6 connectivity or causing packet-handling anomalies. No CVSS score, CWE classification, KEV listing, or POC has been provided at time of analysis, so severity and exploitability remain unconfirmed.

Out-of-bounds read in DHCP options parsing exposes systems running an Ubuntu-maintained networking component to potential memory disclosure or denial-of-service when processing malformed DHCP packets. The vulnerability arises from a size_t integer underflow in the rx_ip path when handling truncated DHCP options fields, allowing an attacker-controlled DHCP server on the local network segment to trigger reads beyond allocated buffer boundaries. No public exploit code has been identified at time of analysis, and exploitation probability data is unavailable due to the recency of this disclosure.

Null pointer dereferences triggered by out-of-memory conditions affect a compression or codec library packaged by Ubuntu, impacting functions precompute_slide_window, more_comps (after failed calloc), and bi_initialize/alloc. Exploitation of these conditions can cause process crashes, resulting in denial of service for any application that processes attacker-supplied data through the affected library. No active exploitation is confirmed; no CVSSv3 score or CWE classification has been assigned at time of analysis.

Out-of-bounds read in Mongoose's TLS server handshake function mg_tls_server_recv_hello exposes systems to memory disclosure or service disruption via a crafted TLS ClientHello packet. An unauthenticated remote attacker can trigger the vulnerability by supplying a malformed session_id_len field during TLS negotiation, before any authentication occurs. No public exploit has been identified at time of analysis, and the disclosure originates from Ubuntu vendor reporting, suggesting a downstream package audit or fuzzing discovery.

Memory misalignment in the Linux kernel's Broadcom/Cypress WiFi SPI driver (brcmfmac) allows incorrect buffer indexing in the cyw_spi_tx function. The function indexes a uint32_t* buffer using a raw byte offset, causing pointer arithmetic to advance by word-sized (4-byte) units when byte-sized steps were intended. Depending on how this misalignment manifests at runtime, consequences may range from corrupted SPI transmit data to an out-of-bounds memory access within the kernel, potentially causing a kernel panic or memory corruption on affected devices. No public exploit or CISA KEV listing is identified at time of analysis.

RX ring stall in the Linux kernel's ENET (Freescale/NXP i.MX Ethernet) interrupt handler causes a network denial-of-service on affected embedded platforms. When an error occurs during packet reception in ENET_IRQHandler, the RX descriptor is not re-owned (returned to hardware DMA control), causing the hardware to believe all remaining descriptors are still held by software and halting all further packet reception. No public exploit has been identified at time of analysis, and this is not listed in CISA KEV, but the network-reachable trigger path means a remote attacker on an adjacent or routed network could potentially stall the interface by provoking the error condition.

Out-of-bounds read in Mongoose's TLS client handshake parser exposes applications using the library's built-in TLS stack to memory disclosure or crash when connecting to a malicious or attacker-controlled TLS server. The flaw resides in mg_tls_client_recv_hello, where a size_t underflow in the ext_len bounds check allows a fixed-offset OOB read beyond the valid message buffer. Reported via an Ubuntu vendor advisory; no confirmed active exploitation or public exploit code has been identified at time of analysis.

Out-of-bounds read in the Linux kernel's IPv6 Neighbor Discovery Protocol stack via the rx_ndp_na handler allows a network-adjacent attacker to trigger a read beyond allocated memory by sending a crafted Neighbor Advertisement packet with a missing or malformed option length field. Affected systems are those running a vulnerable kernel version with IPv6 enabled. Impact is most likely a kernel crash (denial of service) or potential kernel memory disclosure, depending on memory layout at the time of exploitation. No public exploit and no KEV listing identified at time of analysis.

TLS certificate expiry validation in an Ubuntu component uses a hardcoded date string '2025-01-01' instead of the current system time, causing certificate validity checks to produce incorrect results for any certificate whose notAfter field is evaluated against this static anchor. Affected systems may incorrectly accept expired certificates or reject valid ones depending on implementation logic. No CVSS score, CWE classification, KEV listing, or EPSS data is available at time of analysis.

Out-of-bounds read in Mongoose's TLS stack during Client Hello parsing exposes services built on this embedded networking library to remote memory disclosure or crash. The vulnerable function mg_tls_client_recv_hello fails to properly validate the bounds of the key_share extension - a TLS 1.3 construct - when processing incoming handshake messages, allowing a remote attacker to trigger the flaw without authentication before any TLS session is established. No public exploit has been identified at time of analysis, and CISA KEV listing is absent, but the pre-authentication network exposure and trivial triggering conditions make this a meaningful risk for any Mongoose-based service with TLS enabled.

Cryptographically weak TLS secret generation in Cesanta's Mongoose embedded networking library exposes TLS sessions to potential decryption by network observers. The mg_random() function falls back to the standard C library rand() - a non-cryptographic PRNG with predictable output - when a proper entropy source is unavailable, generating TLS keying material that can be predicted or brute-forced. No public exploit has been identified at time of analysis, and no KEV listing exists, but the structural weakness is deterministic on platforms lacking a secure entropy source.

Privilege escalation in LuminalShine prior to version 26.05.0-rc4 stems from a latent file permission gap where `apps.json`, written by the SYSTEM service to `C:\ProgramData\LuminalShine\config\`, could be overwritten by local users if directory ACLs deviate from Windows defaults. On a default Windows installation this EoP path does not trigger because `BUILTIN\Users` inherits only Read+Execute rights from `C:\ProgramData`; the vulnerability is only exploitable when an administrator has loosened those inherited permissions. No public exploit code has been identified and no active exploitation is confirmed.

Improper origin validation in the TP-Link TL-WR820N V2 router's web management interface allows adjacent unauthenticated attackers to extract device configuration details and sensitive information by sending HTTP requests with an absent or empty Referer header. The router relies on the HTTP Referer header as a request legitimacy check, but fails to enforce its presence or value consistently, allowing the validation to be trivially bypassed. No public exploit or active exploitation has been identified at time of analysis, though the attack requires only network adjacency and no credentials.

Denial-of-service in OpenStack Designate's mDNS handler allows unauthenticated remote disruption of DNS resolution across all tenants sharing an affected pool path. The handler performs pool-blind zone lookups: when two zones with the same name exist across different pools - a condition reachable either through a separate cross-tenant zone overlap vulnerability or via legitimate cross-pool configurations - the lookup fails deterministically, causing the handler to return REFUSED for all DNS queries through that path. The _handle_notify trigger is reachable via a single unauthenticated UDP packet, with no public exploit code identified at time of analysis and no confirmed active exploitation in the CISA KEV catalog.

Race condition in IBM DataPower Gateway across three concurrent version lines (10.5.0.x, 10.6.0.x, and 11.0.0.x) permits state contamination of the built-in X-Client-IP header under concurrent request load, enabling IP spoofing and cross-client IP address disclosure. The flaw resides in IBM's enterprise API gateway and integration middleware, commonly deployed in high-throughput service meshes where simultaneous request processing is routine. IBM has released a patch via advisory 7282770, and no public exploit has been identified at time of analysis.

Denial of service in open-iscsi's iscsiuio daemon allows an unauthenticated attacker on the same local network segment to render the daemon permanently unresponsive by sending a single crafted ICMPv6 Router Advertisement carrying a zero-length option, triggering an infinite loop and sustained CPU exhaustion. Affected deployments include Red Hat Enterprise Linux 9 and 10 systems where iscsiuio is active for iSCSI hardware offload. A secondary out-of-bounds read risk exists when a short IPv6 payload is processed, though no memory corruption or data exposure has been confirmed from available intelligence; no public exploit code or CISA KEV listing has been identified at time of analysis.

Cross-site scripting bypass in Loofah 2.25.0-2.25.1 allows attackers to smuggle `javascript:` URIs past the `allowed_uri?` helper by inserting HTML5 named whitespace character references (`	` or `
`) into the scheme. This is a partial bypass of the earlier fix for GHSA-46fp-8f5p-pf2m, which patched numeric character references but not named forms. Only callers that invoke `allowed_uri?` directly with HTML-encoded strings are vulnerable - including Action Text 8.2 markdown link validation - while Loofah's default `sanitize()` pipeline is explicitly unaffected. No public exploit identified at time of analysis; vendor-released patch in version 2.25.2.

IBM Db2 on Linux, UNIX, and Windows exposes cleartext database passwords by writing them into diagnostic trace files, affecting versions 11.5.0-11.5.9 and 12.1.0-12.1.5, including DB2 Connect Server. Any low-privilege local OS user with read access to the Db2 trace directory can harvest plaintext credentials without requiring database-level authentication or elevated system rights. IBM has released a patch; no active exploitation has been confirmed (not in CISA KEV) and no public exploit code has been identified at time of analysis.

Credential interception in Red Hat Multicluster Engine for Kubernetes (MCE) exposes newly rotated provider credentials through an authorization bypass in the provider-credential-controller component. An attacker with specific hub cluster permissions and prior knowledge of a credential value can manipulate `copiedFrom` labels to hijack the credential rotation process, gaining unauthorized read access to sensitive credentials scoped across managed clusters. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis, but the high confidentiality impact and cross-cluster scope change make this a meaningful risk in mixed-trust MCE deployments.

Regex metacharacter injection in @astrojs/netlify prior to 8.1.2 causes the Netlify Image CDN allowlist to be broader than the developer intended, allowing optimization of images from unintended URL paths on already-allowed hostnames. The root cause is that remotePatternToRegex() escapes dots in hostnames but passes literal pathname values into the generated regex unescaped, so a pathname like /img/v1.0/file silently becomes a wildcard that matches /img/v1X0/file or /img/v1/0/file. No active exploitation has been confirmed (not in CISA KEV), though a working proof-of-concept is included in the GHSA advisory; CVSS 3.7 with AC:H reflects the constrained but network-accessible impact.

Plaintext credential exposure in MongoDB SQL Schema Builder CLI allows local users to read authentication material from diagnostic output or log files. The CLI writes its startup configuration - including connection settings containing authentication credentials - to standard output and optionally to a log file on disk without redacting sensitive values. Any local user with read access to the terminal session, the log directory, or a downstream log aggregation system can harvest those credentials and use them to authenticate to the connected MongoDB instance.

IBM i versions 7.3 through 7.6 expose sensitive information to remote authenticated attackers via a race condition (CWE-362). The flaw allows a low-privileged network-authenticated user to read data outside their authorization boundary during a timing window in an undisclosed subsystem. Vendor-released patch is available; no public exploit code exists and the vulnerability is not listed in CISA KEV.

Remote authenticated users can crash applications built on Amazon aws-sdk-cpp versions before 1.11.862 by supplying crafted Base64-encoded input that triggers an out-of-bounds read in the SDK's Base64 decoder. Exploitation is constrained to authenticated users, requires high attack complexity, and is platform-specific - not all deployments are vulnerable. The CVSS 4.0 vector (VA:H, VC:N, VI:N) confirms impact is limited to availability (denial of service); no public exploit code exists and the vulnerability is not listed in the CISA KEV catalog at time of analysis.

Incomplete tilde-escaping in Shescape's Dash shell mode allows attacker-controlled input to bypass sanitization and trigger home-directory expansion. Shescape versions prior to 2.1.14 (v2 branch) and 3.0.1 (v3 branch) on Unix systems fail to escape `~` when it follows `:` or `=`, characters that Dash recognizes as tilde-expansion triggers in assignment contexts. An attacker supplying values such as `:~` can disclose the server process's home-directory path and potentially redirect the filesystem location on which the downstream command operates. No public exploit identified at time of analysis, though a functional proof-of-concept is documented in the vendor's security advisory GHSA-q53c-4prm-w95q.

Filesystem existence and readability probing in Budibase Cloud's MongoDB integration prior to version 3.40.1 exposes an information-disclosure oracle exploitable by authenticated builders. The `tlsCertificateKeyFile` and `tlsCAFile` fields in the MongoDB datasource configuration were passed unsanitized to the MongoDB driver as server-side filesystem paths, enabling a builder to probe arbitrary absolute paths on the shared multi-tenant server via `/api/datasources/verify` by comparing distinct driver error messages. No CISA KEV listing exists, but the GHSA advisory contains a fully working reproduction command; the vulnerability is specific to Budibase Cloud and does not affect self-hosted deployments.

PDF annotation content in SiYuan through v3.7.2 is exposed without authorization via the /api/asset/getFileAnnotation endpoint, which applies only the generic CheckAuth middleware rather than the publish-access controls enforced on the /assets/* route. When publish authentication is disabled, anonymous readers can retrieve .sya annotation files containing private highlights and notes from password-protected, unpublished, or publish-forbidden documents simply by knowing a valid asset path; authenticated publish RoleReaders face the same missing check. The CVSS 4.0 score of 6.9 reflects a network-accessible, unauthenticated vector with limited scope - encrypted notebooks are not affected - and no public exploit code or CISA KEV listing has been identified at time of analysis.

Information disclosure in SiYuan before v3.7.4 allows unauthenticated remote attackers to retrieve the server's absolute filesystem workspace path, operating-system username, and installation layout by querying the resolveAssetPath endpoint with relative asset paths harvested from published documents. The endpoint enforces only a lightweight CheckAuth gate rather than full authentication, satisfying CWE-639's authorization bypass pattern, and returns raw absolute paths to any network-reachable caller. Vendor patch v3.7.4 has been released; no active exploitation has been confirmed and no public proof-of-concept code has been identified at time of analysis.

Unauthenticated information disclosure in SiYuan before v3.7.4 exposes encrypted notebook identifiers, names, and memory-resident unlock states via the getEncryptedNotebookStatus API endpoint without enforcing publish-access filtering. Any anonymous or publish-mode client that can reach the instance can enumerate the full list of encrypted notebooks and determine which are actively decrypted in server memory, leaking sensitive organizational metadata. No public exploit has been identified and no CISA KEV listing exists; a vendor-released patch is available in v3.7.4.

SiYuan's /api/tag/getTag endpoint leaks tag labels and occurrence counts from password-protected documents to any unauthenticated network caller, constituting an authorization bypass under CWE-863. All SiYuan versions before 3.7.4 are affected, as confirmed by a vendor GitHub Security Advisory (GHSA-mp7r-57w4-5qm3) and the EUVD entry EUVD-2026-57452. No public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog; however, the zero-friction network exploitability (PR:N, AC:L, UI:N) means any attacker with network reach can enumerate protected document vocabulary with a single API call.

Information disclosure in SiYuan v3.7.4-alpha.1 (pre-release only) permits any caller who passes basic authentication to retrieve complete database view metadata - including view names, icons, layout types, and per-field Hidden flags - for arbitrary database blocks by supplying an avID, regardless of their access level. The `/api/av/getAttributeViewFieldViews` endpoint introduced in commit acfc02ee8 applies `CheckAuth` middleware but omits any ownership or publish-access filter, producing a CWE-862 Missing Authorization condition. The vulnerability was corrected in the v3.7.4 stable release; no public exploit or active exploitation has been identified at time of analysis.

Unauthenticated information disclosure in SiYuan before v3.7.4 exposes notebook metadata - including names, document counts, storage sizes, and timestamps - for closed and non-published notebooks via the /api/notebook/getNotebookInfo endpoint, which performs no authorization check. Any attacker with network access to the SiYuan API can enumerate structural information that the application's access control model explicitly intends to hide from readers. No public exploit or CISA KEV listing has been identified at time of analysis, but the zero-friction, unauthenticated attack path makes this trivially exploitable wherever the API is internet- or network-exposed.

Information disclosure in SiYuan before v3.7.4 exposes administrator workspace state to unauthenticated network attackers via an improperly protected getConf API endpoint. The UILayout filter fails to enforce authorization boundaries between publish readers and the administrative configuration layer, leaking open document names, recent search terms, notebook directory paths, and private asset locations. No public exploit code or CISA KEV listing exists at time of analysis; however, the unauthenticated network vector makes the flaw trivially reachable on any internet-exposed SiYuan instance with publish features active.

Incorrect privilege assignment in GitLab Enterprise Edition allows a user with pending group membership to receive permissions granted by a custom role, bypassing the membership approval gate entirely. The flaw spans all GitLab EE versions from 15.6 through 19.2.1 - a window of approximately four years - and is documented via a public HackerOne report, though no active exploitation has been recorded in the CISA KEV catalog. The vendor-assigned CVSS score of 3.3 accurately reflects the narrow exploitation conditions: high attack complexity and high privilege prerequisites constrain real-world impact to partial confidentiality and integrity effects within a single affected group.

Integer underflow in GStreamer's avidemux element allows a remote unauthenticated attacker to trigger heap out-of-bounds read and write, heap memory disclosure, and application crash by delivering a crafted AVI file containing a malformed FUJIFILM strd chunk. The gst_avi_demux_parse_strd() function decrements its remaining-length counter by fixed offsets without bounds checking, causing unsigned integer wraparound for payloads of exactly 106 or 107 bytes. Because avidemux is auto-plugged by GStreamer's high-level components (playbin, decodebin, gst-discoverer), any application that opens or previews a media file - including GNOME Files, Nautilus, Totem, and many others - can trigger the vulnerability without user awareness. No public exploit or active exploitation (CISA KEV) has been identified at time of analysis; a vendor-released patch is available in gst-plugins-good 1.28.6.

Out-of-bounds read in GStreamer's avidemux plugin (gst-plugins-good) allows a remote, unauthenticated attacker to crash media-processing applications and potentially leak memory contents by supplying a crafted AVI file. The flaw resides in gst_avi_demux_riff_parse_vprp(), where the count of video field descriptors is derived by dividing remaining buffer size by an attacker-controlled vprp->fields value rather than the struct size, causing the parser to treat more descriptors as valid than the buffer holds. User interaction is required to trigger the bug (the victim must open or stream the malicious AVI via playbin or decodebin), and the primary confirmed impact is denial of service; limited information disclosure is also possible. No public exploit or CISA KEV listing has been identified at time of analysis.

OAuth2 access and refresh token theft in Budibase's automation testing pipeline exposes SSO-authenticated builder credentials to any co-builder of the same application, in all versions prior to 3.39.25. When an SSO-authenticated user runs an automation test in the Budibase builder, the full trigger outputs - including the user's OAuth2 access and refresh tokens - are broadcast without sanitization via WebSocket to every builder connected to the same application room, and are simultaneously stored in a server-side in-memory cache accessible to any co-builder via the test status polling endpoint. A vendor-released patch is available in version 3.39.25; no active exploitation has been confirmed and no public proof-of-concept has been identified at time of analysis.

Account enumeration and targeted denial-of-service in Budibase prior to 3.39.25 stem from an observable response discrepancy in the login lockout mechanism: the login endpoint returns distinct HTTP 403 headers (X-Account-Locked: 1, Retry-After: 900) for locked existing accounts but a generic 403 for non-existent addresses, allowing an unauthenticated remote attacker to verify whether any email is registered in a tenant. Beyond reconnaissance, an attacker can deliberately trigger 5 failed attempts against a known valid address to lock that account for 900 seconds. No public exploit code has been identified and the vulnerability is not listed in CISA KEV; remediation via upgrade to 3.39.25 is straightforward.

TLS certificate verification is silently disabled in the kumactl CLI (Kong Mesh / Kuma) when an operator registers an HTTPS control plane profile without specifying a CA certificate via --ca-cert-file, causing API tokens to traverse the connection without server identity validation. A network-adjacent attacker capable of intercepting traffic between the operator's workstation and the control plane can steal user or administrator API tokens and subsequently issue arbitrary authenticated commands against the control plane. No active exploitation confirmed (not in CISA KEV) and no public exploit code has been identified; patch versions are confirmed by the vendor advisory GHSA-v95x-xhq5-4929.

Kong Mesh's control plane (kuma-cp) exposes the admin bootstrap token and JWT signing keys to any malicious webpage visited by an operator while the control plane is browser-reachable, due to two compounding default misconfigurations. The default CORS wildcard (`CorsAllowedDomains: [".*"]`) reflects any attacker-controlled `Origin` header, and the default `LocalhostIsAdmin: true` promotes loopback-sourced browser requests to `mesh-system:admin` without credential validation - a cross-origin `fetch()` from an attacker page is sufficient to receive full administrative credentials. This affects kuma-cp versions prior to 2.7.25, 2.9.15, 2.11.13, and 2.12.10; no public exploit has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.

Unauthenticated exposure of Envoy admin data in kuma-dp (Kong Mesh data plane) allows any pod on the Kubernetes cluster network to retrieve internal service mesh configuration without credentials. The readiness service on TCP port 9902, bound to all interfaces, acted as an unprotected reverse proxy forwarding nearly all Envoy admin read endpoints - including config dumps, cluster and listener lists, stats, and the mesh trust bundle - to any adjacent network caller. No public exploit code or CISA KEV listing exists at time of analysis; exploitation is bounded by the adjacent network prerequisite, but the risk is elevated in multi-tenant clusters where untrusted pods share the pod network.

Remote authenticated exploitation of a race condition in IBM i 7.3 through 7.6 allows attackers with low-privilege credentials to obtain sensitive information and bypass platform security restrictions. The CVSS vector (AV:N/AC:L/PR:L/UI:N) indicates the flaw is reachable over the network without user interaction, though authentication is required. No public exploit code or CISA KEV listing has been identified at time of analysis; IBM has released a patch via its support portal.

Insecure Direct Object Reference (IDOR) in Apache Allura before 1.19.1 allows unauthorized access to multiple Artifact types due to missing server-side permission checks. Attackers who can interact with the platform can reference artifact objects directly by their identifiers and retrieve content beyond their authorization level, resulting in information disclosure across potentially broad artifact categories. No public exploit code has been identified at time of analysis, and this vulnerability is not listed in the CISA KEV catalog.

Secret exposure in Apache Airflow's Rendered Templates UI allows any authenticated user with task-level view access to read plaintext credential values that should have been masked. The secrets masker's `isinstance(str)` guard silently skips dict-typed Python objects, meaning JSON Variables accessed via `var.json` and containing dict payloads (e.g., database credentials, API keys stored as structured JSON) are rendered in cleartext rather than redacted. No public exploit is identified at time of analysis; however, the disclosure requires nothing more than standard UI navigation by a legitimately authenticated user.

Sensitive values stored as deeply nested lists, tuples, or sets in Apache Airflow Variables are displayed unmasked in the Variables UI due to a recursion-depth logic error in the secrets masker. This is an incomplete-fix follow-up to CVE-2026-42358: that prior fix made dictionary traversal unbounded but left list/tuple/set structures subject to the depth cutoff, allowing values beyond that depth to appear in plaintext on screen. The practical impact is narrow - any authenticated user who can see the Variable in the UI already has read access to the full value through the Variables REST API - making this a shoulder-surfing defense gap rather than an access-control bypass. No public exploit has been identified at time of analysis.

Sensitive credential exposure in Apache Airflow 3.3.0 allows an authenticated Viewer with configuration-read access to retrieve team-scoped sensitive options - including Celery broker URLs with embedded credentials - in cleartext via the Config API. The secrets masker evaluated sensitivity against raw section names without resolving team-prefixed sections (e.g., [team_a=celery]) back to their registered base section, so team-scoped overrides silently bypassed masking while equivalent global options were correctly redacted. This is a third distinct masker bypass variant following CVE-2026-48828 and CVE-2026-48892, meaning deployments that already upgraded to 3.3.0 to remediate those issues remain exposed. No public exploit identified at time of analysis.

Cleartext secret exposure in Apache Airflow's audit log affects all versions before 3.3.1, allowing any authenticated user with audit-log read access to recover Variable values and Connection `extra` credentials verbatim - without holding Variables or Connections read permissions. The bulk API endpoints (`PATCH /api/v2/variables`, `PATCH /api/v2/connections`) bypassed audit-log masking because the masker inspected only top-level request fields while bulk payloads nest entities two levels deep under `actions[].entities[]`. The Airflow UI's Import Variables action posts to this same bulk endpoint, meaning a routine operator import silently writes every secret in the uploaded file to the unmasked audit log. No public exploit is identified at time of analysis, but the vulnerability is fully exploitable by any authenticated user with log access, and deployments that upgraded to 3.3.0 in response to the related CVE-2026-50204 remain vulnerable and must upgrade again.

Cleartext exposure of list-shaped Variable secrets in Apache Airflow Task SDK (all versions before 3.3.1) allows any authenticated Airflow user to recover sensitive values directly from task execution logs and the Rendered Templates UI. The masking logic in _mask_and_deserialize_variable correctly handled string- and dict-typed JSON Variable values but contained no branch for the list type, so arrays - which may hold passwords, API tokens, or connection strings - were returned unredacted. No public exploit has been identified at time of analysis; however, organizations that patched CVE-2026-59244 remain fully exposed because that fix addressed only the dict case, and must upgrade again to 3.3.1.

Arbitrary file write in Progress Software's WhatsUp Gold (all versions before 2026.0.2) enables a privileged attacker with adjacent network access to place attacker-controlled files in web-accessible directories on the host server - a primitive that commonly leads to web shell deployment and subsequent unauthenticated remote code execution via HTTP. Exploitation is constrained by both a high-privilege credential requirement and an adjacent-network attack vector (CVSS:3.1/AV:A/PR:H), substantially narrowing the practical attack surface compared to the full C:H/I:H/A:H impact triad would suggest. No public exploit code or CISA KEV listing has been identified at time of analysis; a vendor-released patch is confirmed available in version 2026.0.2.

FleetDM Fleet's target search endpoint (POST /api/latest/fleet/targets) fails to apply the same field-masking logic used by other team-facing API endpoints, exposing unmasked team enroll secrets and credential-bearing agent options to any authenticated Observer, Observer+, or Technician user before v4.87.0. The practical blast radius extends well beyond information disclosure: leaked enroll secrets allow unauthorized hosts to be enrolled into affected teams, and if agent options contain embedded AWS secret access keys, proxy passwords, or session tokens, those credentials are also retrieved in cleartext. No public exploit code has been identified at time of analysis, though the vulnerability is trivially reproducible by any low-privileged Fleet user.

SQL LIKE wildcard injection in phpMyFAQ's chat user search endpoint (GET /api/chat/users?q=) allows any authenticated user to bypass the intended display-name filter and enumerate all active user accounts. The vulnerability exists because the Chat::searchUsers() method applies database escape() to prevent SQL string breakout but omits escaping of the LIKE metacharacters % and _, which the database engine still interprets as wildcards. A PoC is publicly available demonstrating full user enumeration via q=%25%25 and q=__; no KEV listing exists at time of analysis.

Self-hosted JFrog Artifactory deployments using Helm may expose generated TLS private keys within rendered Kubernetes manifests, where key material persists in cleartext accessible to highly privileged local users. Affecting Artifactory versions below 7.161.16 in the 7.161.x branch and below 7.146.35 in earlier LTS branches, this CWE-312 flaw can facilitate TLS impersonation or decryption of encrypted client-to-repository communications if exposed keys are not rotated after discovery. No public exploit has been identified at time of analysis, and exploitation is structurally constrained by a requirement for pre-existing high-privilege local Kubernetes cluster access.

Insufficient session expiration in JFrog Artifactory (all versions per CPE wildcard) permits an attacker who holds valid credentials for a recently deleted user account to continue authenticating for a brief window after account removal. Classified under CWE-613, the flaw arises from credential invalidation logic that does not immediately propagate account deletion state under specific triggering conditions, leaving a short-lived but exploitable gap. No public exploit code has been identified and the vulnerability is absent from the CISA KEV catalog, but the network-accessible vector means any party with knowledge of the deleted account's credentials could attempt opportunistic access during this window.

The Autodesk Installer IPC frame parser fails to properly validate attacker-controlled position and offset fields (CWE-1285), allowing a locally authenticated low-privileged user to trigger an out-of-range substring operation that crashes the NT AUTHORITY\SYSTEM service. Affected scope covers all tracked versions per the wildcard CPE string, and Autodesk has confirmed and patched the issue via advisory ADSK-SA-2026-0013. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis.

Object Lock COMPLIANCE retention enforcement in RustFS fails open when bucket metadata is absent, unreadable, or corrupt - treating any metadata error as 'no lock configured' rather than blocking the operation. All RustFS versions prior to 1.0.0-rc.1 are affected, exposing COMPLIANCE-retained objects to unauthorized deletion via explicit delete calls, delete_prefix operations, and automated lifecycle or scanner sweeps. No public exploit or CISA KEV listing exists at time of analysis; the risk is primarily to organizations relying on COMPLIANCE-mode immutability for regulatory or legal-hold data retention obligations.

Unauthenticated access control bypass in Fortinet FortiWeb - spanning the 7.0 line through 8.0.2 - and multiple FortiOS branches stems from an incomplete list of disallowed inputs (CWE-184), allowing remote attackers to circumvent input-filtering enforcement and achieve unauthorized integrity modifications. The vulnerability extends to FortiOS, implicating shared blocklist logic across Fortinet's platform stack according to EUVD-2026-57192 and CPE data. Publicly available exploit code exists (CVSS temporal E:P), an official fix has been released (RL:O/RC:C), and no active exploitation has been confirmed in the CISA KEV catalog at time of analysis.

Email injection in the Estatik Real Estate Plugin for WordPress (versions before 4.3.4) allows unauthenticated remote users to send emails to arbitrary recipients with fully attacker-controlled subject, body, and Reply-To headers by abusing the plugin's unvalidated property request form recipient list. The site's own mail infrastructure is effectively weaponized as an open relay, enabling phishing and spam distribution from a trusted sending domain. A publicly available proof-of-concept is referenced in the WPScan advisory; active exploitation has not been confirmed by CISA KEV, though the low skill barrier introduced by the PoC elevates practical risk above what the CVSS 3.7 base score alone suggests.

Fraudulent order completion in the Quick Paypal Payments WordPress plugin (all versions through 5.7.50) enables any buyer to acquire full-price goods or services by paying an arbitrary small amount. The plugin's PayPal IPN handler validates only the order token when determining whether to mark an order as paid - ignoring the payment amount, receiver address, and payment status entirely - so any buyer who can initiate a checkout can trivially complete a full-value order for pennies. Publicly available exploit code exists, reported by WPScan, making immediate opportunistic exploitation straightforward for any motivated attacker targeting affected WordPress storefronts.

Price manipulation in the Payment Button for PayPal WordPress plugin (all versions through 1.2.3.44) enables unauthenticated remote visitors to complete real PayPal transactions at self-specified amounts, causing direct financial loss to the merchant. The plugin forwards whatever payment amount the browser submits to PayPal without any server-side comparison against the merchant-configured price, meaning a buyer can purchase any listed product or service for an arbitrarily low value - including fractions of a cent. A publicly available proof-of-concept exploit exists per WPScan intelligence, lowering the bar for opportunistic financial fraud against any affected storefront.

Payment fraud in the Payment Gateway for PayPal on WooCommerce WordPress plugin (all versions before 9.2.1) enables unauthenticated remote attackers to mark arbitrary WooCommerce orders as paid without completing actual payment. The plugin's PayPal return handler reads attacker-controlled URL parameters and completes orders regardless of whether server-side gateway verification passes or fails, with no amount comparison and no order-ownership validation. A publicly available proof-of-concept exploit exists per WPScan; while the official CVSS score is 5.3, the financial fraud impact on affected e-commerce stores is materially higher in practice.

Payment callback fraud in the Welcart e-Commerce WordPress plugin (versions before 2.11.33) allows any unauthenticated remote attacker to flip their own unpaid order to a settled/paid state by submitting a crafted HTTP request containing only an order number and a status flag - no signature, payment amount, or origin IP is validated. Merchants offering convenience-store or bank-transfer pay-later methods are directly exposed to financial fraud, enabling attackers to obtain goods or services without transferring funds. A publicly available proof-of-concept exists; no active exploitation has been confirmed by CISA KEV at time of analysis.

Authenticated customers on WooCommerce stores running Wallet System for WooCommerce before 2.7.10 can complete purchases for free by submitting arbitrarily inflated wallet deduction amounts during checkout that the plugin accepts without verifying against the customer's actual stored balance. This business logic flaw enables systematic payment fraud - any registered shopper can zero out an order total and receive goods or services without paying the merchant. A publicly available proof-of-concept exists per WPScan, making this trivially exploitable by any site visitor who can register an account, and patching should be treated as urgent.

Cross-team credential disclosure in the Apache Airflow Google Provider's Cloud Secret Manager backend exposes every team's Connections and Variables to every other team in multi-team deployments. The backend's public methods accepted a team_name parameter but silently discarded it before the internal secret lookup, collapsing all team namespaces into one shared, unscoped lookup. No public exploit has been identified at time of analysis; the vendor-released fix is apache-airflow-providers-google 22.3.0.

CSV formula injection in Velociraptor's export pipeline allows attacker-controlled endpoint artifacts to execute as formulas when a forensic analyst opens the exported CSV in Microsoft Excel. Affected export paths include the GUI, offline collector, and data export features - any of which may produce unsanitized CSV output when collected data contains leading formula-trigger characters (=, +, -, @, tab). No public exploit code has been identified and no CISA KEV listing exists; however, the CVSS scope-change metric (S:C) and high confidentiality rating (C:H) correctly reflect that successful exploitation results in formula execution within the analyst's own workstation environment rather than the Velociraptor server itself.

Unauthenticated information disclosure in the Order Sync with Zendesk for WooCommerce WordPress plugin (all versions before 2.2.3) exposes full customer order history and cumulative purchase totals to any remote attacker who supplies a target's email address. The plugin's REST API endpoint omits both a WordPress capability check and any ownership verification, meaning the access control layer is entirely absent - not misconfigured but simply missing. A publicly available proof-of-concept exists (reported by WPScan), and SSVC assesses the vulnerability as automatable, making bulk enumeration of customer purchase data a realistic threat; however, no confirmed active exploitation (CISA KEV) has been recorded at time of analysis.

WPC Admin Columns WordPress plugin exposes arbitrary metadata to subscriber-level authenticated users through an AJAX endpoint that lacks any authorization check. Any account with the lowest default WordPress role - subscriber - can query user, post, and term metadata site-wide, including sensitive data belonging to administrator accounts. Publicly available exploit code documented by WPScan lowers the practical barrier to exploitation; patch version 2.3.4 is available and should be applied immediately.

Eventin WordPress plugin versions before 4.1.20 exposes stored customer records - including names and email addresses - to any authenticated WordPress user holding contributor-level access or higher, due to missing server-side capability checks on customer data retrieval. A publicly available proof-of-concept exploit exists per the WPScan advisory, and an EPSS score of 15% places exploitation probability well above the median for medium-severity CVEs. The vendor-released patch (version 4.1.20) resolves the access control gap; no active exploitation is confirmed in the CISA KEV catalog at time of analysis.

Unauthorized access to paid course content in LearnPress WordPress plugin (all versions before 4.4.4) allows any authenticated user - including low-privilege subscribers - to retrieve lesson material from courses they have not purchased or enrolled in. The flaw resides in the plugin's AI-assistant feature, which processes lesson content requests without first verifying the requesting user's enrollment status, exposing an authorization bypass at the content-serving layer. With a publicly available proof-of-concept from WPScan and a 15% EPSS score, this is a realistic and low-barrier threat to any WordPress e-learning deployment running an affected version.

Kernel information disclosure and local denial of service affect Zephyr RTOS on Intel ALH-equipped hardware through an unvalidated stream_id parameter in the Intel ALH digital-audio-interface driver. A local user-mode thread holding explicit ALH device-object access can invoke the dai_get_properties_copy() syscall with an arbitrary stream_id, triggering an out-of-bounds read from the fixed 64-entry alh_handshake_map[] array and leaking one byte of kernel memory per call, along with a computed FIFO register address that can expose kernel address-space layout. No public exploit has been identified at time of analysis; exploitation is constrained to systems built with CONFIG_USERSPACE and where device-object access has been deliberately granted to untrusted threads.

Stored XSS in tablib's HTML export path before version 3.10.0 allows an authenticated attacker to inject arbitrary JavaScript via maliciously named worksheet titles sourced from imported XLSX, ODS, XLS, or YAML files. The export_book method in _html.py previously interpolated dataset titles directly into HTML h3 tags using an f-string, bypassing all escaping; script payloads assigned to the Dataset title attribute execute in a victim's browser when the exported HTML is rendered. No active exploitation confirmed in CISA KEV; vendor-released patch is available as v3.10.0.

In the Linux kernel, the following vulnerability has been resolved: ovl: check access to copy_file_range source with src mounter creds Commit 5dae222a5ff0c ("vfs: allow copy_file_range to copy across devices") allowed filesystems that implement the copy_file_range() f_op to decide if they want to access cross-sb copy from/to the same fs type. The same commit added checks to verify same sb copy for filesystems that implement ->copy_file_range() and do not support cross-sb copy at the time, namely, to ceph, fuse and nfs. The two remaining fs which implement ->copy_file_range(), cifs and overlayfs started to support cross-sb copy from this time. While overlayfs does support cross-sb copy when the two underlying files are on the same base fs, the copy operation on the two real files from two different overalyfs filesystems is performed with the mounter creds of the destination overlayfs and the read permission access hook for the source file was called with the wrong creds. This could cause either deny of access to copy which would otherwise be allowed (e.g. with splice) or allow read access to file which would otherwise be denied. Fix the latter case by explicitly verifying read access to source file with the source overlayfs mounter creds. The former case remains a quirk of cross-sb overlayfs copy, but userspace could fall back to regular copy so no harm done.

In the Linux kernel, the following vulnerability has been resolved: LoongArch: Fix address space mismatch in kexec command line lookup When searching the loaded segments for the "kexec" command line marker, the kexec_load(2) path (file_mode == 0) passes the user-space segment buffer straight to strncmp() through a bogus (char __user *) cast. This dereferences a user pointer in kernel context, which is wrong and is flagged by sparse: arch/loongarch/kernel/machine_kexec.c:84:51: sparse: incorrect type in argument 2 (different address spaces) @@ expected char const * @@ got char [noderef] __user * Here copy the marker-sized prefix of each segment into a small on-stack buffer with copy_from_user() before comparing, and skip segments that fault. The subsequent copy_from_user() that stages the full command line into the safe area is left unchanged.

CryptoPro Secure Disk for BitLocker before v7.7.4 stores initramfs contents in plaintext on disk, allowing a physical attacker to recover cryptographic keys and other secrets without defeating the encryption layer. The vulnerability was publicly disclosed at Black Hat USA 2026 by researcher Burch in a presentation titled 'The Cost of Obscurity,' suggesting the vendor's security model relied on implementation obscurity rather than cryptographic assurance. No public exploit identified at time of analysis; EPSS at 0.17% reflects the physical-access prerequisite that limits widespread automated exploitation.

Protection-mechanism bypass in CPSD CryptoPro Secure Disk for Bitlocker before v7.7.4 lets unsigned code run from temporary filesystems because the product fails to enforce its Integrity Measurement Architecture (IMA) policy across those locations. An attacker able to stage a binary in a tmpfs-style path can execute code that the integrity policy was meant to block, undermining the pre-boot/disk-encryption trust model. The flaw was disclosed in Black Hat USA 2026 research ('The Cost of Obscurity', Burch); no public exploit is identified at time of analysis, EPSS is low (0.17%, 6th percentile), and it is not in CISA KEV.

Boot-stage integrity bypass in CPSD CryptoPro Secure Disk for BitLocker before v7.7.4 allows an attacker who can present a LUKS-encrypted volume during the pre-boot phase to cause bootxsa.efi to skip all CryptoPro file integrity checks, enabling execution of tampered EFI boot components. Publicly disclosed at Black Hat USA 2026 with an accompanying whitepaper, substantially raising the bar for exploit development. No CISA KEV listing and EPSS at the 1st percentile indicate no observed mass exploitation, but the research-quality disclosure significantly reduces attacker effort for targeted campaigns against affected endpoints.

CryptoPro Secure Disk for BitLocker versions before v7.7.4 leaks TPM2.0 unsealing secrets by persisting them in a serialized binary format within unused disk sectors - raw storage areas that survive reboots and are accessible without operating system authentication. An unauthenticated attacker with physical access to the drive can image these sectors, extract the serialized secrets, and reconstruct a compatible TPM environment to unseal the BitLocker volume, effectively defeating the hardware-backed disk encryption the product is designed to enforce. No public exploit code has been identified at time of analysis; CISA SSVC rates exploitation as 'none', but the attack's low complexity once physical access is obtained was detailed publicly in a Black Hat USA 2026 research presentation, materially lowering the skill barrier for motivated actors.

Full-disk-encryption bypass in CPSD CryptoPro Secure Disk for BitLocker before v7.7.4 stems from a default TPM PCR sealing policy that ignores the system boot state, letting an attacker with physical access unseal the TPM through an unintended execution path or by relocating the drive/TPM to another hardware platform and recover the BitLocker keys. The flaw was documented in the Black Hat USA 2026 research 'The Cost of Obscurity' by Burch; there is no public exploit identified at time of analysis, it is not on CISA KEV, and EPSS is low at 0.18% (8th percentile), consistent with a physical-access rather than mass-exploitable weakness.

CryptoPro Secure Disk for BitLocker before v7.7.4 silently falls back to plaintext volume mounting when a decryption error occurs, negating the entire confidentiality guarantee that BitLocker is meant to enforce. Disclosed publicly at Black Hat USA 2026 in research titled 'The Cost of Obscurity' (Burch), the flaw demonstrates how improper error handling in encryption middleware can fully undermine the security layer it wraps. No public exploit code has been identified at time of analysis, though the Black Hat presentation and accompanying whitepaper are publicly available, substantially lowering the knowledge barrier for exploitation; EPSS stands at a low 0.17% but may shift as presentation details circulate.

Integrity-check bypass in CPSD CryptoPro Secure Disk for Bitlocker before v7.7.4 lets an attacker defeat the product's file-integrity verification: when a LUKS-encrypted volume is present, the software fails to validate that encryption correctly and then silently skips ALL CryptoPro integrity checks, so tampered files pass as trusted. This undermines the pre-boot/endpoint trust guarantees the product exists to provide, enabling evil-maid-style tampering. There is no public exploit identified at time of analysis, though the flaw is the subject of a Black Hat USA 2026 disclosure ('The Cost of Obscurity'); EPSS is very low (0.11%, 1st percentile) and it is not on CISA KEV.

Brute-force credential attacks against Mira's cloud authentication endpoints are unrestricted due to the complete absence of per-account rate limiting, per-IP throttling, and account lockout controls, as disclosed in CISA ICS Medical Advisory ICSMA-26-223-01. Affected products include Mira Firmware 1.7.1.47 and Mira Android App 4.5.15.4 from Quanovate Tech Inc. (operating as Mira / Mira Care), a healthcare-focused IoT platform. An unauthenticated remote attacker can systematically enumerate credentials at arbitrary scale to gain unauthorized access to user accounts, with no public exploit code or CISA KEV listing identified at time of analysis.

Firmware version spoofing in the Mira cloud API allows authenticated users to submit arbitrary firmware version strings for their own registered devices, bypassing server-side validation entirely. Affected products are Mira Firmware 1.7.1.47 and Mira Android App 4.5.15.4 by Quanovate Tech Inc., operating in a medical IoT context and reported under CISA ICS-CERT advisory ICSMA-26-223-01. An attacker exploiting this flaw can cause the vendor backend to misrecord their device's patch state, suppressing security update prompts and distorting fleet-wide vulnerability analytics - with no public exploit identified at time of analysis.

Session token exposure in the Mira Android app (Quanovate Tech Inc.) leaks live authentication credentials to third-party web properties and referrer logs whenever users trigger in-app WebView flows such as shop redirects. The app appends the active session token as a URL query string parameter and embeds a persistent user identifier in the WebView User-Agent header, both of which are transmitted to any destination or JavaScript context receiving that traffic. Reported by ICS-CERT under medical advisory ICSMA-26-223-01, this vulnerability affects a healthcare-adjacent platform, amplifying the sensitivity of exposed credentials; no public exploit code or CISA KEV listing has been identified at time of analysis.

OpenSSH sshd before version 10.5 fails to enforce the `restrict` keyword in `authorized_keys` against tunnel forwarding, allowing a constrained SSH key holder to establish tun-device tunnels that the administrator explicitly intended to block. The flaw affects all OpenSSH releases prior to 10.5 and is classified as a policy-enforcement bypass with low integrity impact. No active exploitation is confirmed per SSVC (exploitation: none) and the vulnerability is absent from the CISA KEV catalog, placing this firmly in the low-urgency tier despite OpenSSH's near-ubiquitous deployment footprint.

Use-after-free (CWE-416) in OpenSSH's ssh client component affects all releases before version 10.5, triggered when a specific pair of remote-forwarding operations occur concurrently, causing access to freed realloc'd memory. A malicious or compromised SSH server could exploit this race condition against a connecting client to achieve partial information disclosure or limited memory corruption. No public exploit has been identified and CISA SSVC rates exploitation as none with partial technical impact, consistent with the moderate CVSS score of 4.8.