Path traversal through dangling symlinks in anshupriyan CLI Bridge prior to 0.0.2 lets attackers who can drive the write_file MCP tool write arbitrary files outside the configured WORKSPACE_ROOT, an integrity-only impact bounded by the privileges of the user running the server. The flaw sits in resolveSafePath: when fs.realpathSync throws ENOENT because a symlink's target does not yet exist, the fallback validates only the parent directory and returns the symlink path as safe, so the subsequent write follows the link out of the sandbox. The project released a fix in 0.0.2 via PR #3 and commit 284a84c, and the independent assessment rates this local-vector (AV:L), unauthenticated, no-interaction integrity impact; no public exploit code or KEV-listed active exploitation was identified at time of analysis.
Credential disclosure in the Apache Camel camel-http component lets a malicious or compromised remote endpoint harvest OAuth2 bearer tokens and basic/digest credentials by answering a request with an HTTP redirect to a host it controls. The leak occurs because OAuth2ClientConfigurer's request interceptor sits in ProtocolExec, below RedirectExec, so 'Authorization: Bearer' headers are re-attached on every redirect hop, and because basic/digest credentials were registered against an AuthScope that matched any host. Affected are Apache Camel 4.0.0 through 4.18.4 and 4.19.0 through 4.22.0; exploitation requires an endpoint with credentials configured and followRedirects=true (not the default) plus an attacker positioned to influence the target server's responses, and no public exploit code or active exploitation (CISA KEV) has been identified at time of analysis.
XML external entity injection in the Apache Camel XPath component lets unauthenticated remote attackers read files accessible to the Camel process and induce it to make outbound requests to attacker-named hosts when an expression is configured with documentType=InputSource and evaluates XML from an untrusted sender. The flaw is configuration-gated rather than default: Camel's default document type already refuses a DOCTYPE, so stock deployments are not exposed, but a deployment that selected InputSource for streaming reasons unknowingly also selected a weaker parser. Fixed in 4.23.0, 4.18.5 and 4.22.1; no public exploit identified at time of analysis, and the CVSS 7.5 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N) reflects confidentiality impact only.
Netis EAP930 access point firmware (build 1.250905.115454) ships a hidden backdoor in the /cgi-bin/debug CGI endpoint, where the return200 function activates concealed functionality when the SIGNATURE argument is manipulated with a specific value. The defect is remotely reachable over the network but is a constrained threat rather than a broadly exploitable one: the assessed vector (CVSS:3.1/AV:N/AC:H/PR:H/UI:N/S:U/C:H/I:H/A:H) requires a high-privilege context and a non-trivial barrier - knowing or forging the correct SIGNATURE - and the vendor itself rates exploitability as difficult. The CVSS v4 vector carries an E:P (proof-of-concept) exploit maturity rating, but no confirmed active exploitation or CISA KEV listing was identified at time of analysis, and Netis did not respond to the disclosure, so no vendor-released patch has been identified.
Signature-verification bypass in bootc and ostree image-mode deployments: a deployment staged while the container signature policy still accepted unsigned content remains bootable after the policy is hardened, because already-staged deployments are finalized on reboot without being revalidated against the current policy. An attacker who can stage a deployment under a permissive policy can therefore get unsigned OS content finalized and booted later, yielding full host compromise (C:H/I:H/A:H); the assessed vector is AV:L/AC:H with operator/local privilege (PR:L) despite the vendor's PR:N rating, and the flaw affects RHEL 8/9/10 image mode, OpenShift Container Platform 4, and Red Hat Hardened Images. No public exploit code was identified at time of analysis and there is no confirmed active exploitation (not in CISA KEV).
Stored cross-site scripting in IBM Cloud Pak System 2.3.6.0 and 2.3.6.1 allows an authenticated low-privileged user to save JavaScript payloads that execute in another user's browser when the affected page is viewed. The flaw is rated CVSS 7.4 (network-reachable, low privileges, scope change) and requires access to a web-console input field whose value is stored and rendered without proper sanitization or HTML encoding. No public exploit code was identified at time of analysis, and IBM has released a vendor patch.
Red Hat Advanced Cluster Security for Kubernetes (RHACS) Central's SAML assertion consumer service (ACS) endpoint does not cache used assertion or response identifiers, so a single valid signed SAML response can be replayed to mint multiple distinct Central authentication tokens for the same mapped identity. Deployments are only exposed when Central is configured to authenticate via SAML; an attacker must first obtain a legitimately signed, unexpired SAML response - they cannot forge one - which is why the issue is rated CVSS 7.4 with high attack complexity despite requiring no privileges or user interaction. Successful replay yields full compromise of confidentiality and integrity for the impersonated identity; no public exploit code or confirmed active exploitation was identified at time of analysis, and non-SAML authentication configurations are unaffected.
Path traversal in Apache Camel's remote-file consumers - the azure-files scheme named in this CVE plus the FTP/SFTP consumers sharing the RemoteFileConsumer base class - lets a malicious, compromised, or multi-tenant remote file server (or an on-path attacker on the integrity-free FTP control channel) cause Camel to read, delete, or rename files outside the directory it was configured to poll, using the endpoint's own credentials. Retrieval is reachable in the default consumer configuration because remote consumers download every listed file; deletion and renaming additionally require delete=true, move, preMove, moveFailed, or readLock=rename. The jailStartingDirectory option, enabled by default and documented as a boundary, did not constrain this remote operand before the fix, so it offered no protection; triggering the traversal requires control over a listing name returned by the server, and for azure-files specifically the Azure Files REST service does not normally return multi-segment names, making exploitation there impractical. No public exploit code or confirmed active exploitation was identified at time of analysis; vendor patches are available.
Path traversal in Apache Camel's remote-file consumers (including the SMB component's share-walking consumer) lets a directory listing entry that is not a single path segment resolve outside the polled remote directory, so Camel will retrieve, delete, or rename files outside that directory using the endpoint's own credentials. Affected are Camel 4.2.0 through 4.18.5-exclusive and 4.19.0 through 4.22.1-exclusive; file retrieval is reachable in a default consumer configuration, while deletion requires delete=true and renaming requires move, preMove, moveFailed, or readLock=rename. The CVSS 3.1 vector (7.4, AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:N) reflects unauthenticated reachability over the network with a high attack-complexity prerequisite: the remote server - or an on-path attacker rewriting an unprotected FTP LIST response, or a malicious, compromised, multi-tenant, or virtual-filesystem-backed server - must emit a qualifying listing name; for the SMB scheme a conforming SMB2 server will not do so, and no public exploit identified at time of analysis.
Path traversal in Apache Camel's remote-file consumers (the FTP, SFTP and mina-sftp schemes) lets a name reported in a remote server's directory listing resolve outside the directory Camel was configured to poll, so Camel will read, delete or rename files elsewhere using that endpoint's own credentials. Affected are Camel 4.18.0 through 4.18.4 and 4.19.0 through 4.22.0, with retrieval reachable in a default consumer configuration because remote consumers download every listed file (CVSS 3.1 base score 7.4, AV:N/AC:H/PR:N/UI:N, C:H/I:H/A:N). The attacker must influence the LIST/SFTP listing output - via a malicious, compromised or multi-tenant server, a virtual-filesystem-backed server, or an on-path attacker against plaintext FTP whose control channel has no integrity protection - and no public exploit has been identified at time of analysis.
Apache Camel's ftp, ftps and sftp remote-file consumers build each file's remote path by concatenating the polled directory with the name returned in the server's directory listing, without any lexical normalization or containment check, so a listing entry containing parent-directory segments resolves outside the configured directory and is then read, deleted or renamed with the endpoint's own credentials. Affected are Camel 4.0.0 through 4.18.4, 4.19.0 through 4.22.0, and earlier; the 4.14.x stream is end-of-life and unfixed, while patches ship in 4.18.5, 4.22.1 and 4.23.0 (CPE: apache_software_foundation:apache_camel). Retrieval is reachable in a default consumer route, whereas deletion needs delete=true and renaming needs move, preMove, moveFailed or readLock=rename; the crucial caveat is that the jailStartingDirectory option, on by default and documented as confining Camel to the starting directory, does not constrain this remote path, so it cannot be used as a mitigation. There is confirmed no active exploitation tracked (not in CISA KEV) and no public exploit code identified at time of analysis, although the reporter privately demonstrated the issue end to end against an FTP endpoint; CVSS 7.4 with AC:H reflects that the attacker must first control or influence the remote server's LIST response rather than attack Camel directly.
OS command injection on the Remote Management Controller (RMC) of HPE Compute Scale-up Server 3200/3250 and Superdome Flex / Flex 280 platforms allows an attacker who already holds high-privilege credentials on the RMC to run arbitrary operating system commands on the controller. The flaw is rated CVSS 7.2 (High) with a vector of AV:N/AC:L/PR:H/UI:N, so exploitation is remote and low-complexity once the privileged foothold exists - the high-privilege requirement is the principal limiting factor and rules out unauthenticated attacks. No public exploit code or confirmed active exploitation is identified in the available data, though the sensitive nature of the RMC as a system-management plane makes exposure to untrusted networks a significant risk multiplier.
Authenticated MISP users can read sensitive server configuration secrets-including the security salt, Redis connection passwords, and plugin API keys-by supplying a configuration branch name such as Security, MISP, or Plugin to the server settings lookup or settings-edit endpoints in versions before 2.5.48. The flaw bypasses the intended redaction check because branch nodes lack the leaf-level redaction flag, exposing the entire configuration sub-tree to any account that can reach those endpoints. No public exploit code or CISA KEV entry was identified at time of analysis, and the impact is limited to confidentiality; a vendor fix is available via commit 30b849102.
Heap out-of-bounds read in rdesktop through 1.9.0's process_raw_bmpcache() lets a malicious RDP server crash the client by sending a raw bitmap cache order that declares large width, height, and bpp values with an undersized payload, causing memcpy() to read past the receive buffer. The attack is unauthenticated but requires the victim to connect to the hostile server (UI:R/UI:P per assessments), and impact is limited to client-side denial of service with no confidentiality or integrity loss. No public exploit identified at time of analysis; the 7.1 CVSS 4.0 score reflects high availability impact only, making real-world priority moderate-to-low.
Authenticated Control Panel users of Craft CMS 4.0.0 through 4.18.7 and 5.0.0 through 5.10.x can bypass the Twig sandbox allowlist through crafted keys in the column, contains, filterByValue, index, where, or multisort array filters, exposing values such as the securityKey, database credentials, and auth tokens. The flaw is limited to accounts that can author or influence Twig template content or field expressions; no unauthenticated path is identified. No public exploit code or confirmed active exploitation was identified at the time of analysis.
MISP before 2.5.48 discloses proposal (shadow attribute) records to authenticated users whose sharing group is attached to the target attribute or object, because the visibility logic accepted any matching sharing group without confirming that the attribute was actually distributed at sharing-group level. The exposure is read-only - an attacker learns proposed attribute changes they should not see - and requires a valid MISP account plus an existing sharing-group overlap with the targeted data; a separate defect also blocked the owning organisation from viewing proposals on its own org-only attributes. No public exploit identified at time of analysis, and the vendor has released a patch in 2.5.48 (commit 9c5af321f).
Improper authentication (CWE-287) in TP-Link Tapo C210 V3.60 firmware lets an attacker positioned on the same adjacent network retrieve shared credential material and cryptographic metadata by sending unauthenticated DS JSON API requests to the camera, without any credentials or user interaction. The recovered shared credential is then accepted by the Streamd Digest authentication mechanism as the administrative password, so the attacker can establish encrypted media sessions and view the camera's protected streams without ever knowing the device account password. Impact is confined to confidentiality - high credential/crypto disclosure and unauthorized media-session access - with no integrity or availability effect cited; exploitation requires local LAN/Wi-Fi adjacency and is not reachable directly from the internet, and no public exploit code or CISA KEV entry was identified at time of analysis.
Heap out-of-bounds read in GIMP's XCF text-layer parasite parsing allows a crafted .xcf file, opened by a victim, to leak adjacent heap memory into layer text and crash the application, causing information disclosure and denial of service. The issue is local and requires user interaction (CVSS 7.1, AV:L/UI:R/PR:N), and the provided intelligence shows no public exploit or CISA KEV entry; it is a moderate-priority issue rather than an urgent remote threat.
Integer overflow in MikroTik RouterOS firewall SAME-target handling can crash or brick affected routers when an authenticated user with firewall/NAT rule privileges installs a manipulated ipt_SAME rule. The flaw affects RouterOS through 6.40.5, 6.44, and 7.20.8, with fixes in 7.21.5 and 7.23.1. No CISA KEV listing was provided and no public exploit was confirmed at time of analysis; the assessed impact is availability-only.
Email manipulation in Krayin CRM before 2.2.7 allows authenticated users with scope-restricted Mail module access to edit, trash, or permanently delete other users' emails by enumerating sequential numeric IDs against EmailController write actions. The update path also discloses email subject, body, and linked lead and person data, and setting is_draft=0 can re-send the victim's email. The vendor fixed this in 2.2.7; no public exploit was identified at time of analysis, and the issue is moderate priority because exploitation requires a valid non-global account.
Insecure direct object reference in Krayin CRM before 2.2.7 allows an authenticated user with a restricted, per-user data scope to read quotes belonging to other users by querying the quotes search and lead-products endpoints, which fail to apply the per-user scope filter in QuoteController. Successful abuse discloses quote addresses, totals, contact details, and product pricing for any user or lead in the instance. The flaw requires a valid low-privileged account (CVSS v4.0 7.1 as published; independently assessed as CVSS 3.1 AV:N/AC:L/PR:L/UI:N/S:U/C:H/I:N/A:N), is read-only with no integrity or availability impact, and has no public exploit identified at time of analysis. Vendor-released patch: 2.2.7.
Denial of service in rdesktop through 1.9.0 lets a malicious or man-in-the-middled RDP server crash the client by sending a crafted revision-2 bitmap cache order to process_bmpcache2() in orders.c. The CVSS vector indicates the server-side attacker is unauthenticated, but the victim must initiate an RDP connection (UI:P); the outcome is a heap out-of-bounds read and client crash with no code execution, privilege gain, or confirmed data disclosure. No CISA KEV entry or public exploit code was identified at time of analysis, and an upstream fix is proposed in PR 437.
Denial of service in rdesktop through 1.9.0 allows a malicious or compromised RDP server to crash the client by sending a malformed 32 bpp compressed bitmap in a bitmap update PDU or bitmap cache order. The server declares a large bitmap geometry while supplying a truncated compressed body, causing process_plane() in bitmap.c to read past the receive buffer; exploitation requires the victim to initiate an RDP connection to the attacker-controlled endpoint, so user interaction is required despite the unauthenticated remote vector. No public exploit code was identified at time of analysis, and CISA KEV does not list this issue.
SGLang 0.5.22 and earlier, when deployed with the Mooncake prefill/decode disaggregation backend, can have its Prefill scheduler memory exhausted by unauthenticated network-adjacent attackers, resulting in the scheduler being OOM-killed and inference service becoming unavailable. The attacker first reads the Prefill rank address from the exposed /route endpoint, then sends STAGING_RSP messages carrying oversized chunk_idx values under fresh room identifiers, so handle_staging_rsp() allocates memory without any bound or validation. Only PD-disaggregated deployments using Mooncake are exposed - default single-node SGLang installations do not exercise the affected path - and there is no public exploit code or CISA KEV entry identified at time of analysis, though the HIGH availability impact (CVSS 7.1, AV:A/PR:N) makes an internal-fabric attacker able to repeatedly restart-cycle the service.
Unauthenticated decode-registration flooding can exhaust memory in SGLang 0.5.22 deployments that use the Mooncake disaggregation backend in prefill/decode mode, ultimately OOM-killing the Prefill scheduler. The attack requires network-adjacent access to the Prefill rank's ZMQ PULL socket and no credentials, but it does not affect default single-node or non-disaggregated SGLang setups. No public exploit was identified at time of analysis, and exploitation is not confirmed in CISA KEV.
Missing authorization in RisingWave before 3.1.0 allows any authenticated low-privilege SQL user with CREATE privilege on a schema to create posix_fs sources or fs sinks that point at arbitrary server filesystem paths, enabling theft of credentials and keys and creation of files with the RisingWave process's privileges. The vendor CVSS 4.0 score is 7.1 (High); the prerequisite is only a valid SQL account with schema CREATE, and no user interaction or special victim-side configuration is needed in affected releases. RisingWave 3.1.0 fixes this by gating local filesystem connectors behind frontend.unsafe_enable_local_fs_connector, which defaults to false in release builds; no public exploit code or active exploitation is identified in the supplied intelligence.
Stack exhaustion in the Qt SVG module (QtSvg) lets a crafted SVG containing deeply nested elements crash applications that parse or render it, producing a denial of service limited to availability. The flaw affects Qt products using QtSvg and is remotely deliverable, but exploitation requires the target application to load the attacker-supplied SVG (CVSS UI:P / assessed UI:R), with no authentication required per PR:N. Source data shows vendor patches but no public exploit code or KEV confirmation, so no public exploit identified at time of analysis.
Reflected cross-site scripting in the Ays Pro Quiz Maker WordPress plugin through version 6.7.1.91 lets unauthenticated remote attackers lure a victim into opening a crafted URL so injected script runs in the victim's browser within the WordPress site origin. The CVSS 3.1 score is 7.1, and exploitation requires user interaction rather than a background attack; the authoritative assessment rates confidentiality and integrity impact as low with no availability impact. No public exploit or CISA KEV entry was identified in the provided data at time of analysis.
Missing authentication in the Remote Control Server component of eremef bandcamp-player through 1.8.33-beta allows unauthenticated attackers with LAN reachability to control the desktop player's remote interface without a device pairing or approval step. The flaw is rooted in how the remoteEnabled setting is handled in database.ts, and it is practically exposed only when the Remote Control feature is switched on and the legacy, non-secure connection mode (plain HTTP/WS, no pairing) is in use; no victim interaction is required against such a server. No public exploit code or active exploitation has been confirmed, and the vendor has released a patched version, 1.8.34-beta, which turns Secure Mode (HTTPS/WSS with QR pairing and desktop approval) on by default.
Unsafe reflective class loading in openGauss-connector-jdbc's LogFactory lets an attacker who controls the JDBC 'logger' connection parameter (URL or connection properties) force the driver to instantiate an arbitrary class present on the client classpath via its String-argument constructor before any Log subtype check occurs, enabling unintended class-loading and constructor side effects. Affected builds run from JDBC 1.1.0 through the 2.x/3.x/5.x/6.x lines (including 6.0.6) and 7.0.0-RC1/RC2/RC3; the flaw is fixed in 7.0.0-LTS. Exploitation is precondition-heavy - local/config-level influence over connection settings (CVSS 7.0, AV:L/AC:H/PR:L) plus a suitable gadget class on the classpath - and no public exploit has been identified at time of analysis.
Sensitive information disclosure in Dell Application Store versions prior to 26.05.0.7 allows an authenticated low-privileged local attacker to read data that should be protected because a critical resource has incorrect permissions. Exploitation is local, requires a non-trivial condition, and is assessed as low urgency; no public exploit identified at time of analysis. Dell's fixed release is 26.05.0.7, documented in DSA-2026-411.
Heap out-of-bounds read in libtins's DNS name parser through version 4.6 can be triggered by a crafted DNS label that ends exactly at the packet buffer boundary, causing a one-byte read past the buffer when additional, answer, or authority records are parsed. The flaw is network-reachable and unauthenticated per the CVSS vector (PR:N/UI:N), but impact is limited to availability: affected applications may crash, especially on hardened allocators, while non-hardened allocators may only read adjacent memory. No public exploit identified at time of analysis.
A regular-expression denial-of-service (ReDoS) flaw in DiffSK configobj through version 5.0.9 lets remote attackers stall Python applications that parse untrusted configuration text, because the _sectionmarker routine in configobj/__init__.py builds a regex vulnerable to catastrophic backtracking. Impact is confined to availability (low-severity availability loss, no confidentiality or integrity effect); unauthenticated attackers can trigger it whenever attacker-influenced input reaches the parser. No vendor patch has been identified at time of analysis - the issue was filed upstream (GitHub issue #281) but the project has not responded, and no public exploit code was identified at time of analysis. Exploitation is highly context-dependent: typical deployments that only parse local, operator-controlled config files expose no remote surface.
Saving attacker-controlled images as GIF in Python Pillow 9.2.0 through 12.x can drive quadratic comment-header assembly and consume excessive CPU, delaying or denying service. The issue is reachable without authentication or user interaction (assessed CVSS:3.1 AV:N/AC:L/PR:N/UI:N), but only when an application preserves a very large embedded GIF comment and re-encodes the image as GIF; impact is limited to availability. No public exploit code or CISA KEV entry was identified in the supplied intelligence, and the vendor-released fix is Pillow 13.0.0.
Denial of service in Pillow 5.4.0 through pre-13.0.0 allows unauthenticated remote attackers to exhaust CPU by supplying a crafted GIF containing numerous comment extensions; Image.open() parses frame zero and triggers quadratic bytes copying in GifImageFile._seek() before the application can inspect the image. Impact is limited to availability (vendor CVSS 4.0 score 6.9; assessed CVSS 3.1 vector AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L), with no code execution or data exposure. No public exploit code or confirmed active exploitation was identified at time of analysis, and environments that only open trusted images are not practically exposed.
Webkul Aureus ERP through 1.6.0 fails to enforce sufficient password-strength requirements in the user creation/edit form implemented in UserForm.php, allowing accounts to be configured with weak passwords. Remote attackers without prior authentication (PR:N) can exploit this weakness through online password guessing or brute-force attempts, although exploitation is considered difficult (CVSS AC:H) and the direct impact is limited to low confidentiality. Publicly available exploit code exists, but there is no confirmed active exploitation (not in CISA KEV), and the CVSS 4.0 score is 2.9.
High-privilege authenticated administrators of Koozali SME Server 11.0rc1-215 can inject reserved Apache configuration directives through the domain creation function, corrupting httpd.conf so Apache fails to start after a subsequent reboot and taking down the web server and hosted virtual hosts. The issue is an input-validation failure (CWE-20) scored 6.9 Medium; exploitation requires administrator access to the management interface and a reboot to trigger the DoS, and no public exploit identified at time of analysis.
Cryptographically weak pseudo-random number generation in the PHP CMS Diem's dmString::random() helper can produce predictable values that may undermine tokens or identifiers derived from it. All rolling-release versions through commit 116974edfb9a5b8bd69cb13586dc62bcdbb485ad are affected; the flaw is remotely reachable by unauthenticated attackers but rated high complexity (CVSS 2.9), with impact limited to low confidentiality. Publicly available exploit code exists, and the vendor has not responded or released a fixed version.
Regular-expression denial of service in Pillow before 13.0.0 lets unauthenticated remote attackers stall an image-processing worker by submitting a tiny crafted XBM file that triggers catastrophic backtracking in the XBM header parser before image dimensions are ever validated. Any service that routes untrusted uploads through Image.open() with XBM parsing reachable is affected; impact is confined to availability degradation of the decoding worker, with no code execution, data exposure, or integrity impact. The flaw is fixed in Pillow 13.0.0, which replaces the backtracking regex with line-based parsing; no public exploit tooling or confirmed active exploitation was identified at time of analysis, though the upstream fix commit includes a regression test that reproduces a minimal triggering input.
Uninitialized stack memory disclosure in Pillow 9.1.0 through 12.x allows an attacker who controls a palette image passed to Image.quantize(palette=p) to receive raw stack contents - including pointer addresses and stack-canary bytes - embedded as pixel indices in the returned PNG or GIF. Exploitation is not implicit in normal Pillow use: the application must expose the quantize-with-palette path to untrusted input and must accept an attacker-supplied palette whose palette data is empty (p.palette.palette == b""), which causes ImagingPaletteCacheUpdate() to leave its local cache array uninitialized before those bytes are emitted by the topalette() conversion path. The flaw yields confidentiality-only impact with no memory corruption or crash; no public exploit code or confirmed active exploitation (CISA KEV) was identified at time of analysis, and the vendor-released fix is Pillow 13.0.0.
FOSSBilling versions 0.1.0 through 0.8.6 leak email-to-client membership through the unauthenticated guest signup endpoint POST /api/guest/client/create: submitting an already-registered address returns a distinct, unredacted exception message from Client\Service::clientAlreadyExists instead of the normal success response, which effectively turns signup into an account-enumeration oracle. The CVSS 4.0 base score is 6.9, but the authoritative assessment rates actual impact as low (CVSS 3.1: AV:N/AC:L/PR:N/UI:N/C:L/I:N/A:N) because it only confirms that an email belongs to a client - no credentials, session, or account data are exposed. No public exploit code or active exploitation is known at time of analysis, and because guest registration can be disabled and pre-0.8.7 builds had no per-email throttling, unthrottled bulk enumeration was the realistic abuse case.
Improper certificate validation in the Google Analytics Client bundled with the Diem PHP framework (dmCorePlugin/lib/vendor/gapi/gapi.class.php) allows an attacker positioned on the network path between a Diem server and Google's Analytics endpoint to intercept or tamper with outbound analytics traffic without detection. Unauthenticated remote attackers can exploit this, but the attack requires a man-in-the-middle position and the vendor itself describes exploitability as difficult (CVSS 4.0 AC:H); impact is limited to low integrity of analytics data with no availability or confidentiality effect. Publicly available exploit code exists (referenced in diem-project/diem issue #454), and no vendor-released patch has been identified because the project has not responded and ships as a rolling release without versioned artifacts.
Path traversal in the TickTick Android app (com.ticktick.task) 8.2.2.0 lets a malicious co-resident application write files to attacker-chosen locations using TickTick's own application privileges. The attacker must already have code installed on the same device and register a ContentProvider whose _display_name value contains traversal sequences, then deliver an ACTION_SEND intent with a crafted content:// URI to TickTick's exported ShareReceiveActivity; the unsanitized name passes through FileManager.getDisplayNameFromUri and AttachmentFileHelper.copyFileDirectly, yielding arbitrary-location writes with high integrity impact but no confidentiality or availability impact. No public exploit code or confirmed active exploitation was identified at time of analysis, and the impact is limited to local, same-device scenarios requiring the victim to have already installed the attacker's app.
VOCALOID6 Editor fails to disable XML external entity resolution when it parses project files, so a crafted project can cause unintended outbound communication, disclosure of files readable by the editor process, or a denial-of-service crash. The issue is rated CVSS 7.1 and requires local file loading plus user interaction (AV:L/UI:R, PR:N), which limits practical exploitation to social-engineering delivery of a malicious project file. No public exploit code or active exploitation was identified in the provided intelligence, making this a genuine but not emergency-priority risk.
Unauthenticated remote attackers can exploit Parse Server user-update handling to confirm guessed usernames and email addresses and, when passwordPolicy.maxPasswordHistory or doNotAllowUsername is configured, validate guessed current or previous passwords while bypassing account lockout. The flaw results from target-account checks being evaluated before caller authorization; affected releases are prior to 8.6.99 and 9.0.0 through 9.10.3-alpha.2. No public exploit code has been identified at time of analysis, and the primary impact is credential enumeration rather than direct compromise.
In aiohttp client versions prior to 3.14.4, a malicious or compromised server can return a Set-Cookie header with a Max-Age value so large that Python's floating-point range is exceeded, raising an OverflowError inside CookieJar.update_cookies; because the exception path failed to close the already-received response, the underlying connection is leaked instead of returned to the pool, temporarily exhausting TCPConnector slots until garbage collection reclaims them. Only applications using aiohttp as an HTTP client with a cookie jar (the default) that fetch from attacker-influenced endpoints are affected, and the effect is availability-only and self-resolving. No public exploit code or confirmed active exploitation was identified at time of analysis; this is a moderate, routine-patching issue rather than an emergency.
An unauthenticated path-traversal flaw in FOSSBilling 0.1.0 through 0.8.6 turns the guest API endpoint api/guest/system/template_exists into a filesystem existence oracle: a remote attacker can pass '..' sequences in the file parameter and read the returned boolean to learn whether arbitrary files or directories exist on the server. No credentials or user interaction are required, but the endpoint never returns file contents, so the practical impact is reconnaissance-grade information disclosure (confirmed file/directory layout) rather than data theft or code execution - which is why the reporter's CVSS 4.0 score of 6.9 sits above the independent assessment's moderate CVSS 3.1 rating (C:L only). No public exploit code was identified at time of analysis and there is no CISA KEV listing; the issue is fixed in 0.8.7.
Server-side request forgery in Crawl4AI 0.8.7 through 0.9.5 lets unauthenticated, network-positioned attackers abuse the Docker API server's fetch_ssl_certificate=true path to make the service connect to internal hosts on TCP/443 and leak TLS certificate metadata or infer reachability. The vendor CVSS 4.0 score is 6.9, while the independent assessment is CVSS 3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:N, reflecting low confidentiality impact only. No public exploit identified at time of analysis.
Unauthenticated information disclosure in Flarum Extension Manager 2.0.0-beta.1 through 2.0.0-beta.7 allows remote attackers to read task output from /api/extension-manager-tasks, exposing filesystem paths, installed extensions and versions, and PHP version details. The TaskResource index endpoint lacks authentication and admin authorization checks (CWE-862), so guests can retrieve data that should be admin-only. The vendor fixed this in 2.0.0-beta.8; no public exploit was identified at time of analysis.
Stack-based buffer overflow in libaribb24 through 1.0.3 allows memory corruption when FFmpeg or VLC decodes ARIB STD-B24 caption data containing a CSI introducer followed by more than 256 parameter bytes. The assessed vector is local with passive user interaction, so exploitation requires a victim to open or process a crafted media file or broadcast stream in a deployment that enables ARIB caption decoding; impact includes crashes and limited integrity effects with high availability loss. No public exploit identified at time of analysis, making this a medium-severity issue rather than an emergency.