Local information disclosure in Apple iOS and iPadOS (and visionOS) before version 27 allows an app running on the device to read sensitive user data it should not be able to reach. Apple fixed the flaw through improved file handling in iOS 27, iPadOS 27 and visionOS 27. No public exploit code or active exploitation has been reported, and the EPSS score of 0.15% (5th percentile) plus SSVC 'exploitation: none' point to a low-volume, user-assisted local attack rather than a mass-exploited one.
An improper authorization flaw in Apple iOS/iPadOS and macOS Tahoe allows a locally installed application to reach sensitive user data that the platform's authorization state checks should have blocked. Apple addressed the flaw through improved state management, shipping fixes in iOS 26.6, iPadOS 26.6, and macOS Tahoe 26.6 for all versions prior to those releases. The issue is local-only (CVSS:3.1/AV:L/AC:L/PR:L/UI:N, 5.5) with confidentiality impact only; EPSS is 0.15% (5th percentile), CISA SSVC rates exploitation as 'none' and technical impact as 'partial', and no public exploit code was identified at time of analysis.
Improper access control in Apple's Keychain service lets a locally installed app delete credentials stored by other applications on iOS 27, iPadOS 27, macOS Golden Gate 27, and visionOS 27 (all versions prior to these releases). The bug is fixed by removing the vulnerable code path; it is rated CVSS 5.5 with integrity-only impact, and there is no public exploit or CISA KEV confirmation at time of analysis (EPSS 0.15%).
Double-free memory corruption in Apple macOS (all versions prior to macOS Golden Gate 27) is reachable by a locally installed or executed untrusted application, which can then trigger unexpected system termination and a local denial of service; confidentiality and integrity are unaffected and there is no remote or network-triggered path. Exploitation requires the attacker to already be able to run code as an application on the target Mac (CVSS:3.1/AV:L/AC:L/PR:L/UI:N, i.e. local, low-privilege context), which is the primary limiting factor against real-world abuse; no public exploit identified at time of analysis and EPSS is very low at 0.14% (4th percentile).
Unauthorized local network access on Apple macOS (all releases prior to Golden Gate 27) lets a malicious or compromised application communicate with local network devices without triggering the macOS Local Network consent prompt. The root cause is improper permission state management (CWE-284), and exploitation is strictly local: the user must launch the offending app (CVSS UI:R), with no remote or unauthenticated path. The independent assessment downgrades confidentiality impact to Low relative to the published CVSS 5.5, limiting practical consequences to reconnaissance of LAN-adjacent devices; no public exploit code has been identified and EPSS is only 0.14% (4th percentile), consistent with a low-priority, targeted issue. Apple has fixed the flaw in macOS Golden Gate 27.
Improper state management in Apple macOS (all versions prior to macOS 27 "Golden Gate") lets a locally installed application reach sensitive user data that normal macOS privacy and access controls are supposed to protect. The attack is local and needs the victim to launch the malicious app (AV:L, UI:R, PR:N), and the impact is confidentiality-only (C:H/I:N/A:N), which yields a moderate CVSS of 5.5. The issue is fixed in macOS 27 per Apple's advisory, EPSS is very low (0.14%, 3rd percentile), there is no CISA KEV listing, and no public exploit code has been identified.
A local denial-of-service flaw in SSSD (System Security Services Daemon) lets an attacker who can connect to the PAM responder socket crash the responder by sending a crafted protocol request that omits the service item while the pam_app_services option is enabled, triggering a NULL pointer dereference. Only availability is affected (CVSS 5.5, C:N/I:N/A:H), but a crash of the PAM responder disrupts authentication on the host until the service is restarted. No public exploit code has been identified at time of analysis and the flaw is not listed in CISA KEV; EPSS data was not provided in the source feeds.
Mattermost's access control policy update endpoint fails to enforce role-based authorization boundaries, enabling a channel or team administrator to detach a system-assigned ABAC parent policy by issuing a crafted PUT /api/v4/access_control_policies request with an empty imports list. Affected are Mattermost versions 11.9.0, 11.8.x through 11.8.4, and 11.7.x through 11.7.7, per vendor advisory MMSA-2026-00724. No public exploit identified at time of analysis, and SSVC assessment confirms no known exploitation with partial technical impact.
Memory leak in the flatcc FlatBuffers C compiler (versions 0.6.0-0.6.3) allows remote attackers to exhaust process memory by supplying crafted schema input that triggers repeated calls to fb_clear_parser() without releasing the keyword_index hash table in the root schema instance. The specific omission - the missing fb_name_table_clear() call on P->schema.root_schema_instance.keyword_index - is confirmed by the patch commit. A proof-of-concept exploit has been publicly disclosed (GitHub issue #389), and SSVC marks the attack as automatable with partial technical impact.
Authorization bypass in IBM Cloud Pak for Business Automation allows authenticated remote users to invoke restricted API endpoints by supplying manipulated HTTP headers that the platform fails to adequately validate. Affected version lines include 24.0.0, 24.0.1, 25.0.0, and 26.0.0, each bounded by specific interim fix thresholds. No public exploit code has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog.
IBM Sterling B2B Integrator (6.2.0.0 through 6.2.0.6_2, 6.2.1.0-6.2.1.2, 6.2.2.0-6.2.2.1) and IBM Sterling File Gateway Standard Edition in the same version ranges contain an improper authentication flaw (CWE-287) that lets a remote authenticated attacker bypass security restrictions. The CVSS 3.1 base score is 5.4 (AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:N), indicating a low-privilege authenticated network attacker can gain limited read and modify capability without user interaction, but no confidentiality-integrity-availability loss beyond partial. There is no CISA KEV listing, no public exploit code identified, and no EPSS score supplied in the available intelligence, so this appears to be a targeted-patch issue rather than a mass-exploitation emergency.
Stored cross-site scripting in IBM Cognos Analytics 12.0.4 through 12.0.4 FP2 and 12.1.0 through 12.1.3 FP1 allows any authenticated non-privileged user to embed JavaScript in their own given name or surname profile fields. The payload is later rendered without encoding in the administrative user-account management panel, so the script executes in an administrator's browser session when that account's permissions are reviewed, potentially exposing the administrator's session cookies. The issue is rated CVSS 5.4 (medium) and a vendor patch is available; no public exploit code or CISA KEV entry was present in the provided intelligence, so active exploitation is not confirmed.
Stored Cross-Site Scripting in the web backend of Repetico 1.9.7.31 for Android allows an authenticated remote user to inject and execute arbitrary JavaScript in other users' app contexts by embedding a malicious payload in a multiple-choice question text field. The CVSS Scope:Changed metric confirms the injected script can affect victim users beyond the attacker's own session, enabling session hijacking, credential theft, or unauthorized in-app actions. A researcher-published PoC writeup on GitHub confirms the attack path. No vendor patch or advisory has been identified at this time.
Authenticated HTML injection in IBM Sterling Secure Proxy 6.2.0.0 through 6.2.1.2 lets a logged-in user plant unescaped markup in the management interface, so the product renders attacker-controlled content as trusted UI. The result is UI spoofing and phishing: a crafted page can impersonate legitimate Sterling Secure Proxy screens to harvest credentials or trick operators into unsafe actions, with limited confidentiality and integrity impact (CVSS 5.4). No public exploit code has been identified and the flaw is not listed in CISA KEV; the vendor has published a fix, though an exact fixed version is not named in the provided data.
Blind server-side request forgery (SSRF) in IBM WebSphere Application Server 8.5 and 9.0 enables authenticated remote attackers to issue arbitrary internal HTTP requests by embedding malicious URLs within crafted SOAP messages processed by the server. Although the response is not directly returned to the attacker ('blind' SSRF), an adversary can exploit timing discrepancies and error differences to map internal network topology, probe cloud instance metadata endpoints (e.g., AWS IMDSv1), or reach services otherwise inaccessible from the public internet. No public exploit code has been identified and this CVE is not listed in the CISA KEV catalog at time of analysis.
Server-side request forgery in IBM Langflow OSS versions 1.0.0 through 1.11.2 allows authenticated network attackers to induce the server to issue unauthorized HTTP requests to arbitrary destinations, enabling internal network enumeration or use of the server as a proxy to reach otherwise-isolated services. The vulnerability is classified as CWE-918 and is confirmed by IBM with a patch available at the vendor support page. No public exploit code and no CISA KEV listing have been identified at time of analysis.
Reflected cross-site scripting in IBM WebSphere Application Server 8.5 and 9.0 allows a remote unauthenticated attacker to inject and execute arbitrary JavaScript in a victim's browser by tricking them into clicking a specially crafted link to the vulnerable server. Successful exploitation can lead to session token theft, credential harvesting, or unauthorized actions performed under the victim's browser context. No public exploit code or active exploitation has been identified at time of analysis, and a vendor patch is available.
Stored cross-site scripting in the share renderer of Trilium Notes v0.103.0 and earlier lets anyone with note-authoring privileges embed arbitrary JavaScript in a webView note via the unescaped #webViewSrc attribute, which then executes in the browser of every user who opens the shared note — including administrators. The flaw requires an authenticated author to create or modify the note and a victim to open the shared link (CVSS 5.4, PR:L/UI:R/S:C), so it is a privilege-escalation and session-hijack vector rather than an unauthenticated remote compromise. CISA SSVC rates exploitation as none and automatable as no, and no public exploit code or KEV listing is present in the supplied intelligence.
IBM Db2 Mirror for i 7.4, 7.5, and 7.6 accept forged requests without adequate validation, letting a remote attacker ride an authenticated administrator's browser session to read and lightly modify data exposed through the Db2 Mirror interface. Exploitation requires the victim to be logged in and to interact with attacker-controlled content (CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:L/I:L/A:N, base 5.4), so this is a cross-site request forgery rather than a pre-authentication takeover. There is no CISA KEV listing, no public exploit code identified, and CISA's SSVC assessment rates exploitation as 'none' and automatable as 'no'; IBM has published a fix via its support advisory.
An improper authentication flaw (CWE-287) in IBM Langflow OSS versions 1.0.0 through 1.10.2 lets a remote attacker who already holds a valid low-privileged account change the password of another user's account, resulting in account takeover rather than direct code execution. The CVSS 3.1 score is 5.4 (AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:L/A:L), reflecting that exploitation requires an authenticated session and primarily damages integrity and availability of the targeted account. No public exploit code has been identified and the issue is not listed in CISA KEV, but SSVC flags partial technical impact and the vendor has published a fix; the exact patched version string is not enumerated in the available data.
Unrestricted file upload in DotVVM before 4.2.11, 4.3.15, and 5.0.0-preview09-final lets unauthenticated remote users push arbitrary files into the application's configured upload storage by posting directly to DotvvmFileUploadMiddleware without a valid X-DotVVM-UploadToken, enabling repeated uploads that exhaust disk space and cause denial of service. The flaw stems from the middleware storing files without verifying a per-render cryptographic token or applying the previously client-supplied type/size constraints server-side. No CISA KEV listing or public exploit code was identified at time of analysis; the vendor has released fixes and adds an optional AuthorizeFileUpload authorization hook.
Improper resource handling in FFmpeg 8.0.x's HLS playlist parser allows remote attackers to cause denial of service by supplying a crafted HLS playlist containing malformed `duration` or `target_duration` field values. The flaw resides in `parse_playlist()` within `libavformat/hlsproto.c` and requires passive user interaction - the target application must process the attacker-controlled `.m3u8` playlist. No public exploit code or active exploitation (CISA KEV) is known; vendor-confirmed fixes are available in FFmpeg 8.1 and 9.0.
IBM i Navigator for i — the browser-based administration console shipped with IBM i 7.3, 7.4, 7.5 and 7.6 — can be made to ignore the binding that ties an administrative session to the client's source IP address, allowing a remote, unauthenticated attacker (CVSS PR:N/UI:N) to reach sensitive information that should have been confined to a legitimate session. IBM scores the exposure 5.3 with confidentiality-only impact (C:L/I:N/A:N), consistent with a targeted information disclosure rather than a full compromise of the host. Vendor fixes are published, and no public exploit code or CISA KEV entry appears in the supplied intelligence, so this is best treated as a patch-on-schedule hardening item rather than an emergency.
Unsafe deserialization in FedML up to version 0.9.6 allows low-privileged remote attackers to potentially execute arbitrary code by manipulating the `s3_key_str` argument passed to `S3Storage.read_model` in the MQTT+S3 communication backend. The function deserializes model data retrieved from S3 without adequate input validation, meaning an attacker who can influence the S3 key or the object stored at that key can supply a crafted malicious payload. No patch has been released; the vendor has not responded to the coordinated disclosure filed in GitHub issue #2267. No public exploit code or CISA KEV listing has been identified at time of analysis.
Stack-based buffer overflow in MikroTik RouterOS's mtget binary TFTP RRQ builder allows any authenticated user, including those with only read-only group membership, to crash the mtget worker process by issuing a /tool fetch command with a crafted tftp:// URL path of 507 bytes or more. The overflow involves an unbounded rep movsb instruction that overwrites saved registers at a deterministic stack offset, causing a reproducible process crash without requiring network connectivity to an actual TFTP server. No public exploit code has been identified at time of analysis, and vendor patches are available in RouterOS 7.23.4 (long-term) and 7.24.2 (stable).
Missing authorization in the A2A Java SDK (a2aproject/a2a-java) 1.2.0 allows a caller with a valid account to invoke the ListTasks operation without the list-scoped read check that guards other task operations, so task listings are returned to principals that the configured TaskAuthorizationProvider may have intended to reject entirely. The flaw lives in AuthorizationRequestHandlerDecorator.onListTasks in server-common, which delegated to the handler without first consulting the authorization provider for the list scope, leaving protection to per-task filtering inside the TaskStore. It is fixed in release 1.3.0 (tag v1.3.0.Final, commit e9a1abf9c90c02b16d17293afdc3cc2f555d63a6); CVSS 4.0 base is 5.3, there is no CISA KEV listing and no public exploit code identified at time of analysis.
Network traffic integrity can be silently undermined on unpatched Apple platforms - iOS and iPadOS before 27, macOS Golden Gate before 27, tvOS before 27, and watchOS before 27 - by an unauthenticated attacker who already holds a privileged on-path network position between the device and its peer. Apple fixed the underlying cryptographic weakness through improved integrity checks in the 27-generation releases, restoring the guarantee that intercepted traffic cannot be modified undetected. Exploitation is constrained by the adjacency requirement (AV:A) and high attack complexity (AC:H), EPSS is only 0.17% (7th percentile), and no public exploit code has been identified at time of analysis.
DocsGPT through version 0.20.0 leaks OAuth connector session tokens and provider account emails to attacker-controlled windows via an insecure postMessage implementation. The callback-status endpoint broadcasts OAuth tokens using a wildcard target origin ('*') without validating the sender's origin, allowing any page acting as window.opener during the OAuth authorization flow to intercept the message payload. An attacker who tricks a victim into initiating an OAuth cloud storage connector flow from an attacker-controlled page can steal session tokens and subsequently disconnect the victim's cloud storage connectors. No public exploit has been identified at time of analysis; EPSS data not provided.
Server-side request forgery in Refly through version 1.1.0 allows authenticated users to coerce the backend into issuing HTTP requests to arbitrary internal addresses by supplying unvalidated URLs to the POST /v1/misc/scrape endpoint. The backend performs no scheme, hostname, or resolved-IP validation before fetching, enabling access to loopback services, RFC1918 private networks, and link-local cloud metadata services such as AWS IMDS at 169.254.169.254. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis; EPSS data is not yet available for this CVE.
Cross-site scripting in Takahe through 0.11.0 allows any remote ActivityPub actor to inject executable `javascript:` scheme URLs into federated post content or profile summaries. The sanitizer in `core/html.py` does not block this URL scheme when processing inbound federated HTML, so a crafted link persists in the instance's rendered content. When a local user clicks such a link, the JavaScript executes in the Takahe instance's origin, enabling session hijacking or account impersonation without requiring any privileges on the target server. No public exploit or active exploitation has been identified at time of analysis.
IBM WebSphere Application Server 9.0 and 8.5 fail to neutralize CRLF and other control characters coming from remote input before writing them into the server's administrative log, letting an unauthenticated network attacker insert forged log entries (CWE-117, improper output neutralization for logs). The CVSS 3.1 score of 5.3 reflects low integrity impact only - confidentiality and availability are unaffected and no code execution is possible - but falsified administrative logs can mask real activity and mislead incident responders. IBM has published a fix (support page node 7286610); there is no CISA KEV listing and no public exploit identified at time of analysis, though CISA's SSVC assessment marks the issue as automatable.
Shell command injection bypass in HKUDS nanobot 0.2.0-0.2.1 allows authenticated users to evade the ExecTool allow/deny pattern filter by embedding newline or carriage-return characters in command strings. The `_split_shell_segments()` function in `shell.py` treated only `;` and `|` as command delimiters, so a payload on a subsequent line (e.g., `echo allowlisted\nrm -rf /tmp/sensitive`) passes the filter entirely unchecked. A proof-of-concept for this technique is confirmed by SSVC, and an upstream patch commit (af582246f141) is available but no tagged release has been confirmed.
Missing authorization in the Django-CRM bulk case update endpoint (backend/cases/bulk_views.py) allows any authenticated member of an organization to edit, reassign, or close cases belonging to other users. Affected releases are django-crm 1.0 through 1.2; the same code path also skips the model's close-transition validation, so a case can be closed without the recorded approval that the single-case workflow requires. Version 1.3.0 resolves the issue (patch commit 799bb121), CVSS 4.0 rates it 5.3 (AV:N/AC:L/PR:L/UI:N), SSVC records no exploitation and no automatable vector, and no public exploit code has been identified at time of analysis.
Let me synthesize the analysis for CVE-2026-90790, an SSRF in a2a-python's push notification sender. Key facts: - Product: a2a-python (A2A Python SDK by a2aproject) - Function: _dispatch_notification in src/a2a/server/tasks/base_push_notification_sender.py - Component: Push Notification Sender - Argument push_info.url → SSRF (CWE-918) - Remote exploitation possible - Affected: up to 1.1.3 (1.1.0, 1.1.1, 1.1.2, 1.1.3) - Fix: 1.1.4 (release tag v1.1.4, PRs 1164 and 1169) - CVSS 4.0: 5.3, AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:L/VA:L/SC:N/SI:N/SA:N - Note PR:L — low privileges required. So authenticated in some sense. - EPSS: not provided - KEV: no - POC: "Reported by VulDB" — no explicit POC. The patch PRs are public, so the fix commit is public. But no explicit POC listed. I'll say "no public exploit identified at time of analysis". - Tags: Python, SSRF, A2A Python Patch: available from vendor, exact version 1.1.4. Interesting detail from diff: the library removed the built-in `push_url_validation_error` helper (SSRF screening) and made the validator opt-in, requiring the deployment to pass `push_url_validator` explicitly. Actually the diff shows the removal of `_ip_is_blocked` and `push_url_validation_error` from base_push_notification_sender.py, and adding a push_url_validator parameter that defaults to None (no library screening). Hmm — that's interesting: the fix in v1.1.4 removed the automatic SSRF screening and replaced it with opt-in validation via `validate_push_notification_url` in src/a2a/utils/push_url_validator.py. That's a bit odd — the fix may be reworked. Actually it looks like the "fix" makes the URL validation a shared utility and requires explicit opt-in. Hmm, this suggests the vulnerability was that the default screening wasn't applied in `_dispatch_notification` at all (the function didn't call the validator), so pushing to a malicious URL would trigger an outbound request to an arbitrary internal address. The PR adds the validator hook into `_dispat
Cross-tenant sender alias disclosure in Froxlor before 2.3.7 allows any authenticated customer to enumerate and read other customers' allowed sender alias values via an IDOR flaw in the delete-confirmation flow of `customer_email.php`. The query fetching `allowed_sender` uses only the global auto-increment `senderid` parameter without scoping it to the requesting customer, making all sender alias records on the instance accessible to any authenticated tenant. Publicly available exploit code exists, verified against Froxlor 2.3.6 with `mail.enable_allow_sender` enabled; no active exploitation has been identified.
Authorization bypass in Froxlor before 2.3.7 allows authenticated customers to create MySQL databases and users on any operator-configured server, regardless of per-customer server restrictions. The Mysqls.add API command validates the customer-supplied mysql_server parameter only for numeric format and server existence - the calling customer's allowed_mysqlserver allowlist is never consulted when the parameter is explicitly provided in the request. In multi-tenant shared hosting deployments, this enables a customer to access premium-tier, isolated, or cross-tenant database server pools they were explicitly barred from, bypassing the operator's resource segregation controls. SSVC assessment confirms proof-of-concept exists; no active exploitation identified.
The Microsoft Calendar plugin in Mattermost fails to validate null entries in incoming Microsoft Graph webhook payloads, enabling any unauthenticated remote attacker to crash the plugin process by sending a single crafted POST request to the public webhook endpoint. Affected are Mattermost versions 11.9.0, 11.8.x through 11.8.4, 11.7.x through 11.7.7, and 10.11.x through 10.11.22. No public exploit code or active exploitation has been identified; however, SSVC classifies this as automatable, meaning the denial-of-service condition can be triggered reliably and at scale without special tooling.
Disk space exhaustion in multer 2.2.0-2.3.0 allows a remote unauthenticated attacker to repeatedly start and abort multipart/form-data uploads, leaving orphaned temporary files that accumulate until the host disk partition is exhausted. The race condition arises when asynchronous disk writes complete after multer's abort cleanup routine has already run, so those files escape deletion entirely. No authentication is required and no special server configuration beyond the default disk storage engine is needed; the issue is fixed in multer 2.4.0.
Stack-based buffer overflow in Contec Remote I/O Coupler Unit (Server Type) CPSN-MCB271-* allows an authenticated remote attacker to trigger a denial-of-service condition by sending a specially crafted network request. The overflow (CWE-121) corrupts stack memory, crashing the device firmware without evidence of code execution capability. No public exploit or KEV listing exists; JPCERT/CC published the advisory coordinated with Contec.
Out-of-bounds write in CONPROSYS nano Series Remote I/O Coupler Units from Contec Co., Ltd. enables remote attackers with low-privilege network access to crash affected industrial automation devices by sending specially crafted network requests. Affected hardware includes server-type couplers (CPSN-MCB271), EtherNet/IP adapter variants (CPSN-EOB471EI), and software PLC-type units (CPSN-PCB271-S1-041), placing operational technology environments at risk of physical process disruption from device-level DoS. No active exploitation has been confirmed and no public proof-of-concept code is known at time of analysis.
Directory listing exposure in CONTEC CONPROSYS PAC Series enables unauthenticated network attackers to enumerate internal directory structures by requesting a specific URL on the device's embedded web server. Affected models include the Integrated Type CPS-PC341 and Configurable Type CPS-PCS341, both deployed in industrial automation environments. No public exploit has been identified at time of analysis, and no CISA KEV listing exists; impact is limited to information disclosure that could assist subsequent reconnaissance against the device.
Directory listing exposure in Contec's CONPROSYS M2M Gateway and M2M Controller Series allows a remote attacker to enumerate server-side directory structures by accessing a specific URL on the device's web interface. The affected hardware includes both integrated and configurable variants of the CPS-MG341, CPS-MGS341, CPS-MC341, and CPS-MCS341 product lines. Exposed directory listings reveal file and path structures that can facilitate reconnaissance for follow-on attacks against these industrial gateway devices. No public exploit or active exploitation has been identified at time of analysis.
Reflected cross-site scripting in regularlabs.com's Articles Anywhere (versions before 20.0.0) and Users Anywhere (versions before 2.1.0) extensions for Joomla allows unauthenticated remote attackers to inject executable JavaScript into victims' browsers. The extensions expose a data-tag templating system that pulls values from visitor-controlled request parameters and renders them in output without context-sensitive encoding - Joomla's built-in string input filter does not distinguish between safe-for-HTML-text, safe-for-attribute, and safe-for-URL contexts, so a malicious request value can materialize as an event handler attribute or a javascript: URL wherever a site author has placed a data tag. No public exploit identified at time of analysis.
An out-of-bounds write (buffer overflow) in the PASE AIX-runtime layer of IBM i 7.3, 7.4, 7.5, and 7.6 lets an authenticated local user crash their own PASE process, producing a self-limited denial of service. The independently assessed vector (CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:L) requires a valid IBM i account able to execute code in PASE, and per the vendor the impact is confined to terminating the attacker's own process rather than affecting other users, sessions, or the underlying operating system; the vendor's NVD entry scores it 5.2 under a scope-changed (S:C, I:L) interpretation. EPSS is 0.12% (2nd percentile), there is no CISA KEV listing, and no public exploit code has been identified at time of analysis, so real-world risk is low and this should not be triaged as a critical memory-corruption event despite the buffer-overflow framing.
An out-of-bounds write in Apple's operating system memory handling lets a process already confined to the platform sandbox break out of those restrictions, gaining partial read and write access to data it should not reach. Affected platforms are iOS and iPadOS before 26.7/27, macOS Sequoia before 15.8, macOS Tahoe before 26.7, macOS Golden Gate before 27, visionOS before 27 and watchOS before 27. Apple has released fixes for every affected platform, and CISA's SSVC assessment records no exploitation, no automation and only partial technical impact.
Stored XSS in MISP ≤2.5.45 allows an authenticated user who can upload SVG files to plant malicious JavaScript that executes on the MISP origin when any other authenticated user navigates directly to the SVG URL (organisation logos or event-report pictures). Because the vulnerability lives in the serve path rather than the upload path, SVG payloads deposited while uploads were enabled remain exploitable even after SVG upload support is later disabled. No public exploit or CISA KEV listing identified; CVSS 4.0 scores this 5.1 and SSVC rates technical impact as partial with no known exploitation.
The YAMAP - Social Trekking GPS App for Android exposes users to information leakage and unauthorized redirects through an improperly controlled WebView (in-app browser) implementation. The vulnerability stems from insufficient access control over which URLs or communication sources the WebView will process, allowing a crafted link or external input to steer the browser toward attacker-controlled content or harvest app-internal data. No active exploitation has been confirmed (not in CISA KEV) and no public exploit code has been identified at the time of analysis, but user interaction is required, making social engineering the realistic attack vector.
Cross-Site Request Forgery in MISP versions ≤2.5.45 allows a remote unauthenticated attacker to silently modify settings of an already-authenticated MISP user by inducing them to load a malicious page. Three UserSettingsController endpoints - setTheme, setHomePage, and eventIndexColumnToggle - were placed on the Security component's unlockedActions list, completely disabling CSRF token and field-hash validation for those POST endpoints. The highest-impact vector is setHomePage, which enables an attacker to redirect a threat analyst's MISP landing page to an attacker-controlled URL, creating a plausible pivot for credential phishing or social engineering; no active exploitation or public proof-of-concept has been identified at time of analysis.
Cross-site scripting in the Image Management component of Contec SolarView Compact (models sv-cpt-mc310 and sv-cpt-mc310f) enables authenticated attackers to inject malicious script that can trigger arbitrary OS command execution on the underlying device. Reported by JPCERT/CC (JVNVU97753461) with a Contec vendor advisory, the vulnerability chains CWE-79 XSS with a command injection pathway, making it more dangerous than its CVSS 4.0 score of 5.1 suggests. No public exploit code and no CISA KEV listing have been identified at time of analysis.
Stored cross-site scripting in Contec SolarView Compact (sv-cpt-mc310 and sv-cpt-mc310f) solar power monitoring appliances affects the Schedule Settings and Mail Send Setting pages of the web management interface. An attacker who injects a malicious script into these configuration fields can cause arbitrary OS command execution when a privileged user renders the poisoned page in an authenticated session. No public exploit has been identified at time of analysis, and the vendor has issued a security advisory coordinated through JPCERT/CC under JVNVU97753461.