Full write control over configured S3 session recording buckets is achievable in Ivanti Endpoint Manager (EPM) before version 2024 SU7 by remote authenticated attackers exploiting a CWE-73 filename control flaw. The Core component fails to sanitize user-supplied input used to construct S3 object paths, enabling any low-privilege authenticated EPM user to write arbitrary data to the session recording bucket - potentially overwriting or corrupting forensic evidence. No public exploit code or CISA KEV listing has been identified at time of analysis; the fix is available as EPM 2024 SU7.
Null pointer dereference in the Windows TCP/IP kernel driver exposes virtually every supported Windows Server and desktop release to remote denial-of-service attacks without any authentication or user interaction. The flaw, rooted in CWE-476, causes a kernel panic (Blue Screen of Death) when a specially crafted network packet triggers a NULL pointer dereference in tcpip.sys, crashing the affected host. Nineteen distinct product/version combinations from Windows Server 2012 through Windows Server 2025 and Windows 11 25H2 are confirmed affected; no public exploit has been identified at time of analysis, and Microsoft has released corrective patches.
Sandbox escape in Cursor IDE for macOS prior to version 3.1.2 allows an AI agent operating in Auto-Run Sandbox mode to replace a Python virtual environment executable with a malicious wrapper, which the Microsoft Python extension then invokes outside the sandbox boundary - granting arbitrary host command execution with the user's full macOS privileges, including modifying files outside the workspace and launching applications. The vulnerability exploits a privileged trust relationship between the sandboxed agent and an unsandboxed host-level extension. No active exploitation has been confirmed (not in CISA KEV), and no public exploit code has been identified at time of analysis; a vendor-released patch is available in version 3.1.2.
Token credential exposure in Red Hat insights-client causes the CCX_TOKEN bearer credential, used exclusively in disconnected cluster deployments, to be written in clear text to pod logs on every startup when glog verbosity is set to level 2 or higher. Any party with read access to pod logs or an aggregated logging backend can silently harvest this credential and use it to authenticate against the CCX API, extending access beyond the originating pod (CVSS scope change S:C). No public exploit is identified at time of analysis, and EPSS at 0.21% (11th percentile) combined with SSVC exploitation status of 'none' indicates this remains theoretical rather than actively targeted.
Arbitrary file write in n8n's Edit Image node allows authenticated workflow users to escape the node's designated working directory by injecting unsanitized format parameter values into the underlying image library. All n8n releases before 1.123.67 on the 1.x line, before 2.31.5 on the 2.x line, and before 2.32.1 on the 2.32.x line are affected, as confirmed by the GHSA-xmc9-4f2h-jf9c advisory from the n8n maintainers. No public exploit code has been identified at time of analysis, but patched versions and a remediation commit are available, and no special attacker position beyond workflow execution privileges is required.
Arbitrary code execution in Adobe Lightroom Classic versions up to 15.4.1 is achievable when a victim opens a specially crafted malicious file, exploiting an Incorrect Authorization flaw (CWE-863) that allows attacker-controlled code to execute in the context of the current user. The attack scope extends beyond the application boundary (S:C), amplifying impact to potentially affect the underlying OS or co-running user-context processes. No public exploit or CISA KEV listing has been identified at time of analysis, but the file-based delivery model and creative professional user base make this a targeted phishing risk.
Remote code memory disclosure in the Windows Remote Desktop Client (CWE-125) exposes sensitive process memory to an attacker operating a malicious or compromised RDP server. All major Windows desktop and server releases from Windows 10 1607 through Windows 11 26H1 and Windows Server 2012 through Windows Server 2025 are affected, representing essentially the entire active Windows install base running an RDP client. Microsoft has released patches via the MSRC advisory; no public exploit code or CISA KEV listing exists at time of analysis, and SSVC rates exploitation status as none.
Remote information disclosure in the Windows SMB Client affects every actively-supported Windows 10, Windows 11, and Windows Server release from 2016 through 2025, exposing potentially sensitive memory contents to unauthenticated network attackers via a CWE-125 out-of-bounds read (CVSS 7.5). The flaw resides in the SMB client stack and is triggered when a vulnerable Windows host processes a crafted response from an attacker-controlled SMB server, requiring no authentication or user interaction from the attacker's perspective. No public exploit has been identified at time of analysis and CISA has not added this to the Known Exploited Vulnerabilities catalog; however, the extraordinary breadth of affected products and the vendor-confirmed patch make this a standard patch-cycle priority.
Command injection in Visual Studio Code versions prior to 1.132.1 enables remote unauthenticated attackers to disclose sensitive information from a victim's system over a network. Exploitation requires user interaction - the victim must open or engage with attacker-controlled content such as a malicious workspace, configuration file, or repository - consistent with the CVSS UI:R metric. No public exploit code or CISA KEV listing is identified at time of analysis; a vendor-released patch (1.132.1) is available via the Microsoft Security Response Center.
Windows iSCSI Target Service across multiple Windows Server and Windows 10 versions contains a null pointer dereference that enables an unauthenticated remote attacker to crash the service, resulting in denial of service. Affected systems span Windows Server 2016 through 2025 and Windows 10 versions 1607 and 1809, all below their respective patched build numbers. No public exploit code has been identified at time of analysis, and the vulnerability is not listed in the CISA KEV catalog, though the low-complexity, unauthenticated network vector makes it straightforward to exploit once the target service is reachable.
Incorrect Authorization in Adobe Commerce and Magento Open Source allows a low-privileged authenticated attacker to bypass security controls over the network without user interaction, gaining unauthorized read access and high-integrity write access across affected store resources. The vulnerability covers a wide span of actively supported release lines - Adobe Commerce 2.4.4 through 2.4.9, Magento Open Source 2.4.6 through 2.4.9, and Adobe Commerce B2B 1.3.3 through 1.5.3 - making the exposure broad for merchants who have not applied August 2026 security patches. No public exploit code or CISA KEV listing has been identified at time of analysis, but the high integrity impact (I:H) is particularly concerning for e-commerce environments where unauthorized write access could affect orders, pricing, or customer records.
Out-of-bounds read in the Windows Remote Desktop Client component (CWE-125) exposes high-sensitivity memory contents to remote network attackers, enabling information disclosure across all supported Windows 10, Windows 11, and Windows Server 2012 through 2025 releases. The vulnerability carries a CVSS 7.5 base score with a network attack vector and no privilege requirement per the official vector, though SSVC classifies exploitation as not automatable and technical impact as partial, suggesting real-world exploitation is constrained by the need for a user to initiate an RDP connection to an attacker-controlled server. No public exploit code has been identified at time of analysis, EPSS places active exploitation probability at 0.85% (55th percentile), and a vendor-released patch is available via Microsoft MSRC.
Typebot's Google Sheets OAuth integration exposes live workspace-level bearer tokens to any guest member via the `getAccessToken` helper, which checked only for read-level workspace membership before decrypting and returning the raw credential. Versions prior to 3.17.0 allow a low-privileged guest to enumerate credential identifiers, invoke the endpoint, and reuse the workspace's Google OAuth token entirely outside Typebot to access or manipulate connected Google Sheets. No CISA KEV listing exists at time of analysis, but SSVC classifies exploitation status as proof-of-concept with network attack vector and low privilege required, making this a realistic threat for any Typebot workspace with Google Sheets integrations.
Credential exfiltration in Typebot prior to 3.17.0 allows a low-privilege guest workspace member to steal decrypted OpenAI-compatible API keys by invoking the model-listing endpoint with an attacker-controlled baseUrl. The server decrypts the stored credential and forwards the secret in both the Authorization header and an explicit api-key header to whatever URL the guest supplies, effectively acting as a credential-delivery proxy. This vulnerability has no public exploit identified at time of analysis, but the attack is trivially exploitable by any authenticated workspace guest with read access.
Path traversal in SAP Manufacturing Integration and Intelligence allows a highly privileged attacker to write files outside intended directory boundaries by supplying specially crafted input to affected file path validation functions. Exploitation requires a legitimate user to subsequently interact with the attacker-influenced content and depends on conditions outside the attacker's direct control, as reflected by the AC:H and UI:R CVSS metrics. Successful exploitation achieves a scope change affecting adjacent components with high impact on confidentiality, integrity, and availability. No public exploit code and no CISA KEV listing exist at time of analysis.
Information disclosure in Windows Presentation Foundation (WPF) within .NET 8, 9, and 10 allows an unauthenticated network attacker to exfiltrate sensitive data when a target user opens a specially crafted document. The WPF document parser resolves external references from attacker-controlled URIs - an SSRF-adjacent primitive (CWE-829, CWE-918) - causing outbound network requests that can expose authentication tokens or internal metadata to the attacker's infrastructure. No public exploit code or CISA KEV listing exists at time of analysis; vendor-released patches are available for all affected .NET and Visual Studio versions.
Denial-of-service via out-of-bounds read in the Ivanti Endpoint Manager Agent (all versions before 2024 SU7) allows remote unauthenticated attackers to crash the agent service with a single crafted network request. The CVSS 7.5 vector (AV:N/AC:L/PR:N/UI:N) confirms no authentication or user interaction is required, making this trivially exploitable from any network-reachable position. No public exploit code or CISA KEV listing has been identified at time of analysis; impact is limited to availability disruption of the EPM agent, with no confirmed confidentiality or integrity consequence.
Sensitive memory disclosure in Adobe ColdFusion 2023 and 2025 stems from the server's use of a broken or risky cryptographic algorithm (CWE-327), enabling a remote unauthenticated attacker to retrieve confidential data - such as session tokens, credentials, or application secrets - that transit or reside in vulnerable cryptographic operations. Both the 2023 and 2025 product lines are confirmed affected per Adobe's own CPE data. No user interaction is required, and the CVSS vector confirms full network exploitability against default deployments. No public exploit code or CISA KEV listing has been identified at time of analysis.
Sensitive information disclosure in Microsoft Teams for iOS (versions 2.0.0 through prior to 8.8.1) enables unauthenticated network attackers to access protected data without any user interaction, as reflected in the CVSS:3.1 vector AV:N/AC:L/PR:N/UI:N with C:H. The vulnerability is rooted in CWE-200 improper information exposure and is iOS-platform-specific, confirmed by CPE cpe:2.3:a:microsoft:teams:*:*:*:*:*:iphone_os:*:*. No active exploitation is confirmed - the CVE is absent from the CISA KEV catalog, SSVC assessment rates exploitation as none, and EPSS sits at 0.68% (49th percentile), indicating low observed exploitation probability at time of analysis.
Sandbox escape in Activepieces prior to version 0.80.0 allows authenticated users with Code step creation rights to execute arbitrary Node.js code in the host engine process, bypassing the V8 isolate entirely. The flaw exists because the engine calls importFresh() - a wrapper around Node.js require() - to load user-supplied compiled modules before the V8 sandbox is instantiated, meaning top-level module code runs unrestricted in the host process. An attacker exploiting this can exfiltrate platform-critical environment secrets (AP_ENCRYPTION_KEY, AP_JWT_SECRET), read or write arbitrary files on the host, and reach internal services; no public exploit has been identified at time of analysis, but key compromise would have platform-wide cascading consequences.
Remote unauthenticated denial of service in Windows Network File System (NFS) allows a network attacker to crash the NFS service by sending a malformed packet that triggers a buffer over-read (CWE-126). Affected systems span Windows Server 2012 through 2025 and Windows 10 versions 1607 and 1809. No public exploit code has been identified at time of analysis, and CISA has not added this to the Known Exploited Vulnerabilities catalog; Microsoft has released patches across all affected build lines.
Unauthenticated credential theft in the InstaWP Connect WordPress migration plugin (all versions ≤ 0.1.3.6) exposes database access details and API signatures on Apache servers during active migration windows. The plugin predictably derives its AES-256-CBC decryption passphrase as SHA256(migrate_key), where the migrate_key is directly leaked in the filename of the unprotected options file via Apache directory listing - allowing an attacker who finds the file to mathematically recover the encryption key and decrypt the options payload without any brute force. Recovery of the api_signature may enable downstream remote code execution via InstaWP's service API. No public exploit or CISA KEV listing has been confirmed at time of analysis, though the deterministic key derivation significantly reduces exploitation effort.
Remote code execution via use-after-free in the Windows DNS service allows an authenticated, network-based attacker to execute arbitrary code on affected Windows Server and Windows 10 systems. The vulnerability spans a wide range of Windows Server versions from 2012 through 2025 and is exploitable over the network, though high attack complexity (AC:H) limits opportunistic exploitation. No public exploit identified at time of analysis; vendor-released patches with specific build targets are available via the Microsoft Security Response Center.
Security feature bypass in Microsoft Visual Studio Code Copilot Chat Extension (versions 0.27.0 through pre-1.132.1) enables network-based attackers to circumvent a protective control without authentication, resulting in high integrity impact and low availability impact. The attack requires user interaction - consistent with a scenario where a victim must engage with crafted content delivered over the network. No public exploit code has been identified at time of analysis, and a vendor patch (1.132.1) is available via the Microsoft Security Response Center.
Uncontrolled resource consumption in Adobe's CAI Content Credentials SDK enables remote unauthenticated attackers to exhaust system resources and cause an application denial-of-service. Three distinct SDK components are affected: the Rust SDK (c2pa), the JavaScript/WebAssembly SDK (@contentauth/c2pa-web), and the command-line tool (c2patool), all across their respective versions prior to the patched releases. No public exploit code or active exploitation has been identified at time of analysis, though the CVSS vector (AV:N/AC:L/PR:N/UI:N) indicates this is trivially exploitable without authentication or user interaction.
Null pointer dereference in Adobe's Content Credentials SDKs (Rust SDK ≤c2pa-v0.90.5, JS SDK ≤@contentauth/c2pa-web@0.12.1, CLI tool ≤c2patool-v0.27.5) enables unauthenticated remote attackers to crash any application built on these libraries by supplying a crafted C2PA media payload. The CVSS:3.1 vector (AV:N/AC:L/PR:N/UI:N) confirms straightforward remote exploitation with zero prerequisites or user involvement. No active exploitation has been confirmed in CISA KEV and no public exploit code has been identified at time of analysis; vendor-released patches are available for all three SDK components.
Incorrect Authorization (CWE-863) in Adobe Commerce and Magento Open Source allows remote unauthenticated attackers to bypass security controls and gain unauthorized read access to protected resources. Affected are all major supported branches of Adobe Commerce (2.4.4 through 2.4.9-2026-jul), Adobe Commerce B2B (1.3.3 through 1.5.3-2026-jul), and Magento Open Source (2.4.6 through 2.4.9-2026-jul). No public exploit has been identified at time of analysis, though the CVSS:3.1 AV:N/AC:L/PR:N/UI:N profile makes this trivially exploitable at scale once a technique is known.
Unauthenticated remote directory listing in deepwiki-open through commit 16f35a0 allows any network-accessible attacker to enumerate arbitrary filesystem paths on the host by supplying absolute path parameters to the local-repository structure endpoint. The root cause is a combination of missing input path restriction (CWE-22) and an authentication layer that is disabled by default (WIKI_AUTH_MODE=false), meaning every default deployment is immediately vulnerable without any configuration change. No public exploit has been identified at time of analysis, but the zero-authentication, low-complexity attack surface makes this trivially exploitable against any internet-exposed instance.
Information disclosure in OpenSignLabs OpenSign through 2.37.0 exposes complete organisation tenant records to unauthenticated remote attackers via the gettenant Parse cloud function. The function accepts a contactId parameter and returns the full tenant configuration record without any authentication or authorization check, enabling systematic enumeration of all tenant data across a multi-tenant deployment. No public exploit has been identified at time of analysis, but the trivial attack complexity (AV:N/AC:L/PR:N/UI:N) means no specialised tooling is required.
Unauthenticated contact record enumeration in OpenSignLabs OpenSign through 2.37.0 exposes all stored personally identifiable information to remote attackers without credentials. The `getcontact` Parse cloud function executes with `useMasterKey`, which bypasses Parse Server's native access control layers (ACLs and Class-Level Permissions), and contains no compensating authentication or authorization logic. Any network-reachable attacker can call this function, supply arbitrary object IDs, and retrieve full contact records including PII at scale. No public exploit has been identified at time of analysis, and no KEV listing exists.
Server-side request forgery in Dub's metatags edge endpoint allows unauthenticated remote attackers to direct the server to issue HTTP requests to arbitrary internal or external hosts. All Dub versions through 2026-07-10 are affected, with the root cause being the endpoint's acceptance of caller-supplied URLs without any denylist filtering or authentication requirement. Attackers can leverage this to probe RFC 1918 address space, enumerate internal services, or exfiltrate credentials from cloud metadata endpoints such as AWS IMDSv1; no public exploit code or active exploitation (CISA KEV) has been identified at time of analysis.
Authentication bypass in Red Hat's search-v2-api exposes federated search results to unauthenticated remote attackers. The authentication middleware unconditionally skips validation when an HTTP request carries the `Upgrade: websocket` header, allowing any network-reachable attacker to POST to the `/federated` endpoint without credentials. Exploitation yields information disclosure across all configured remote managed hubs - no active exploitation confirmed (not in CISA KEV) and no public POC identified at time of analysis, but the CVSS 7.5 score and PR:N vector reflect low-barrier, network-accessible exploitation.
Unauthenticated remote write to any contact record is possible in OpenSignLabs OpenSign through version 2.37.0 via the updatecontacttour Parse cloud function, which performs no authentication or authorization checks before accepting and applying updates to arbitrary contact records. Any internet-reachable deployment is vulnerable to data corruption or overwrite of contact records for all users system-wide without credentials. No public exploit code or CISA KEV listing has been identified at time of analysis, though the vulnerability's zero-prerequisites nature and straightforward exploitation model make it a realistic target for automated abuse.
Unauthenticated remote memory exhaustion in Undertow's WebSocket subsystem enables denial of service against any deployed application exposing a Jakarta @ServerEndpoint with an @OnMessage handler. An attacker needs only a standard WebSocket handshake - no credentials, no special tooling - to trigger unbounded Out of Memory conditions on the server. No public exploit code has been identified at time of analysis and the vulnerability is not listed in CISA KEV, but the zero-prerequisite attack vector (CVSS AV:N/AC:L/PR:N/UI:N) makes exploitation trivially automatable against any reachable Undertow instance.
Unauthenticated remote access to all contact form submissions, including personally identifiable information, is possible in CSZ CMS 1.3.2 due to a broken access control flaw in the admin form-submission viewer. The endpoint omits an authentication check entirely, and the underlying framework authentication helper fails open rather than denying access by default - meaning the bypass is a structural failure, not a simple missing decorator. No public exploit code or active exploitation has been identified at time of analysis, but the trivial exploitation conditions (no credentials, no special configuration) make this a realistic data-exposure risk for any internet-facing CSZ CMS deployment.
Unauthenticated remote access to invoice PDFs in IDURAR ERP CRM 4.1.0 and earlier exposes customer PII due to the /download Express router being mounted globally without authentication middleware, constituting a broken access control flaw (CWE-284). Any internet-reachable attacker can retrieve financial documents by enumerating MongoDB ObjectIds, requiring no credentials, session token, or user interaction. No active exploitation has been confirmed via CISA KEV, and no public exploit code is identified in the available intelligence, though the attack requires only basic HTTP requests and ObjectId pattern knowledge.
Unauthenticated exposure of live task execution data in Kestra prior to 2.0.0-rc6 allows any network-accessible party to retrieve plaintext credentials, environment variables, HTTP headers, and connection strings via an unprotected GET /worker endpoint on the worker process. While the main Kestra API on port 8080 enforces authentication, the worker's WorkerEndpoint.java endpoint was not subject to the same controls, serializing the full live Task object to unauthenticated callers. The vendor-released fix is confirmed in version 2.0.0-rc6; no public exploit code has been identified at time of analysis.
Uncontrolled heap allocation in WildFly's CSIv2Util GSS token decoder allows a remote unauthenticated attacker to trigger OutOfMemoryError by supplying a crafted length field, crashing or destabilizing the Java application server. The CVSS vector (AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H) confirms no authentication or user interaction is required, making this a trivially exploitable denial-of-service condition against any WildFly deployment with an accessible IIOP endpoint. No public exploit code has been identified at time of analysis, and this CVE does not appear in the CISA KEV catalog.
Denial of service in Red Hat JBoss Enterprise Application Platform's jboss-remoting component allows any remote unauthenticated attacker who can reach application or management ports to exhaust server memory through a crafted protocol handshake. By completing an Upgrade: jboss-remoting handshake on ports 8080, 9990, or 4447, an attacker triggers an integer overflow (CWE-190) that causes uncontrolled heap allocation, inducing OOM errors that degrade or halt request processing server-wide. No public exploit has been identified at time of analysis and the CVE is not listed in the CISA KEV catalog.
Use-after-free memory corruption in MongoDB Server's $graphLookup aggregation stage allows an authenticated high-privilege user to crash the server or potentially execute unintended code by combining aggregation and memory-management commands in a way that causes a stale internal reference to be dereferenced after the underlying memory is freed. Affected versions are MongoDB Server 8.3.0 through 8.3.7; a vendor-released patch is available in 8.3.8. No public exploit code or CISA KEV listing has been identified at time of analysis, making this a high-severity but access-gated vulnerability.
Resource exhaustion in iperf3 server mode allows unauthenticated remote attackers to trigger a Denial of Service by submitting crafted control-channel JSON with oversized numeric values for the `parallel` and `len` parameters. The server performs no bounds validation on these fields, causing it to spawn excessive threads or streams and allocate disproportionately large buffers until system resources are exhausted. No public exploit code has been identified at time of analysis, and this vulnerability is not currently listed in the CISA KEV catalog.
Unauthenticated account creation in Swing Music 3.0.0 exposes a critical access control gap: the POST /auth/profile/create endpoint is explicitly allowlisted from JWT verification, enabling any network-reachable attacker to register arbitrary user accounts with no credentials. Once registered, the attacker-controlled account can be used to authenticate and access server-protected functionality such as media libraries or administrative features. No public exploit code has been identified and the vulnerability is not listed in the CISA KEV catalog, though the low-complexity network vector makes opportunistic exploitation straightforward against exposed instances.
Server-side request forgery in Pinry through 2.1.13 enables unauthenticated remote attackers to coerce the application server into issuing arbitrary HTTP requests to internal or external hosts via the pin-from-URL feature, which passes attacker-supplied URLs directly to Python's requests.get() without host allowlisting or IP validation. Because ALLOW_NEW_REGISTRATIONS defaults to true, no account or authentication is required to trigger the vulnerable endpoint, making every default-configured internet-exposed Pinry instance exploitable without any attacker preconditions. Attackers can leverage this to reach cloud metadata endpoints such as AWS IMDSv1 at 169.254.169.254, harvest IAM credentials, or probe internal services unreachable from the public internet; no public exploit has been identified at time of analysis and no CISA KEV listing is present.
Denial of service in Red Hat EAP's Undertow HTTP/1.1 chunked-transfer decoder allows unauthenticated remote attackers to exhaust JVM heap memory by sending crafted requests with no chunk size or count limits enforced. The OutOfMemoryError triggered halts all deployments on the affected listener, not just the targeted application. CVSS metrics (AV:N/AC:L/PR:N/UI:N/A:H) confirm trivial exploitability with no authentication and no special conditions, making this a high-priority availability risk for any internet-facing EAP deployment. No public exploit or CISA KEV listing identified at time of analysis.
Missing authorization in calibre's Content Server allows readonly or anonymous users to persist unauthorized annotation changes to library books prior to version 9.12.0. The POST /book-update-annotations endpoint omits the needs_db_write=True decorator flag, causing the router to bypass ctx.check_for_write_access() and pass attacker-controlled JSON directly into the database via db.merge_annotations_for_book(). No public exploit code exists and the vulnerability is absent from CISA KEV, though the trivial one-line root cause lowers the barrier to exploitation on exposed instances.
OpenSign through version 2.37.0 exposes an unauthenticated remote integrity bypass in its document audit-trail mechanism, allowing any network-accessible attacker to fabricate arbitrary audit log entries by directly invoking the triggerevent Parse cloud function with caller-controlled viewer identity and IP address parameters. The vulnerability (CWE-345: Insufficient Verification of Data Authenticity) directly undermines non-repudiation in digitally signed documents, which is the legal and evidentiary foundation of electronic signature workflows. No public exploit has been identified at time of analysis, but the CVSS vector confirms the attack requires no authentication, no user interaction, and low complexity, making it trivially exploitable against any exposed OpenSign deployment.
Use-after-free in Google Chrome's Extensions subsystem prior to version 151.0.7922.137 enables an attacker to execute arbitrary code inside the Chrome sandbox by convincing a target to install a crafted malicious extension. The vulnerability class (CWE-416) indicates improper memory management in the extension handling code, where a freed memory reference is later dereferenced, corrupting heap state in a way an attacker can direct toward code execution. No public exploit or CISA KEV listing is identified at time of analysis; exploitation requires deliberate social engineering to achieve extension installation.
OS command injection in Google Turbinia's forensic worker fleet allows an attacker who can submit processing requests or control evidence path/name values to execute arbitrary commands on the worker nodes. The root cause, confirmed by PR #1631, is multiple workers (grep, photorec, strings) constructing shell command strings via Python format interpolation and invoking them with shell=True, allowing shell metacharacters in evidence paths to break out of the intended command. The fix, released 2026-07-10, replaces all affected workers with list-based subprocess invocation, eliminating shell interpretation.
Out-of-memory denial-of-service in ffuf prior to v2.2.0 allows a malicious HTTP server to crash the fuzzer by returning a decompression bomb - a tiny compressed payload that expands to gigabytes - because the response-size guard in pkg/runner/simple.go checked only the compressed Content-Length while io.ReadAll consumed the fully decompressed stream without any byte cap. Any ffuf user fuzzing a hostile endpoint is affected; no special configuration is required. The issue is fixed in v2.2.0 via io.LimitReader; no public exploit has been identified and it is not listed in CISA KEV.
Privilege escalation via use-after-free in the Windows Kernel affects Windows 11 (24H2, 25H2, 26H1) and Windows Server 2025, requiring physical device access to exploit. An attacker with physical proximity can trigger kernel memory corruption under high-complexity conditions to gain full control of the affected system, achieving complete confidentiality, integrity, and availability impact. No public exploit code has been identified at time of analysis, and CISA SSVC confirms exploitation status as none.