We need to produce a JSON object. All fields are plain text strings. We must synthesize from all provided data. We have CVE-2026-59859, affecting Kiota's PHP generator, code injection via unescaped $ in double-quoted literals. Patch available. EPSS 1.02%, percentile 59%. CVSS vector: 4.0. Need to compute assessed CVSS 3.1 and 4.0 independently based on description. SSVC: Exploitation none, Automatable no, Technical impact total. Tags: PHP, RCE, Code Injection, Kiota. KEV not mentioned, so not in KEV. So no active exploitation confirmed. POC? Not mentioned as public. So no public exploit identified? But note that it's a code generation literal injection; exploit could be crafted spec. I'll follow rule: "no public exploit identified at time of analysis". Patch available from vendor. Fix in Kiota version 1.32.4. NuGet packages: Microsoft.OpenApi.Kiota and Microsoft.OpenApi.Kiota.Builder, both fixed in 1.32.4. So remediation: upgrade to version 1.32.4. Affected products: Microsoft Kiota (CPE cpe:2.3:a:microsoft:kiota:*:*:*:*:*:*:*:*) prior to 1.32.4. References: GitHub advisory, PR, commit, release tag. For summary: synthesize. Not copy description. Start with impact verb and product. E.g. "Code injection via malicious OpenAPI specs in Microsoft Kiota PHP generator < 1.32.4 allows arbitrary PHP code execution." Mention attacker controls OpenAPI spec, generated code contains unescaped $ leading to PHP interpolation. Mention if actively exploited? KEV no, so "not known to be actively exploited". EPSS low (1.02%) suggesting low probability. So summary: "Code injection vulnerability in Microsoft Kiota's PHP code generator allows attackers who control OpenAPI specifications to inject arbitrary PHP code into generated client classes. Versions before 1.32.4 embed schema-derived strings without escaping the dollar sign ($), enabling PHP double-quoted string interpolation at runtime." Mention exploitation conditions: requires attacker to provide a crafted OpenAPI spec, whic
We need to produce a comprehensive JSON analysis. The user asks for a senior vulnerability analyst synthesis across multiple intelligence sources. The provided data includes CVE description, CVSS (4.0 vector), CWE-94, EPSS 1.02% (59th percentile), patch available, CPE, tags (RCE, Code Injection, Kiota), SSVC: exploitation POC, automatable yes, technical impact total. References: GitHub security advisory, PR, commit, release tag, VulDB, NVD, EUVD. Source code evidence from PR diff. We must extract product_name: "Kiota" (Microsoft Kiota, but simply "Kiota" is fine; "Microsoft Kiota" is three words, but the rule says "1-3 words ONLY. The specific product or service name affected. Examples: 'Apache Tomcat', 'WordPress', 'Linux Kernel', 'nginx-ui', 'GoAnywhere MFT'. NOT the vulnerability type, NOT the vendor alone, NOT a description." So product name is "Kiota". Could also be "Microsoft Kiota" (2-3 words). I'll go with "Microsoft Kiota" (2 words) to be precise, but the examples show "Apache Tomcat", "GoAnywhere MFT". So "Microsoft Kiota" is appropriate. That is exactly the product. The vulnerability is in the C# code generator. Affected package: nuget/Microsoft.OpenApi.Kiota and Builder. So product_name: "Microsoft Kiota". That's 2 words, okay. Summary: Need to synthesize, not copy description. Start with impact verb and product. E.g., "Code injection in Microsoft Kiota allows attackers to inject arbitrary C# code into generated HTTP clients." Must include version affected (<1.32.3), what attacker can do. Mention if actively exploited (KEV status isn't provided, but tags include RCE, but KEV is not listed. The input does not have a KEV field. So we should not claim KEV. The SSVC says exploitation: poc, meaning proof-of-concept exists. That's public but not actively exploited per KEV. So I'll state "publicly available exploit code exists" because SSVC says poc and advisory includes an example, and PR has tests showing injection. So summary: "Code injection in Microsoft
Arbitrary JavaScript execution in Microsoft's @prompty/core TypeScript loader (2.0.0-alpha.1 through 2.0.0-beta.2) allows an attacker who supplies a crafted .prompty file to run code during prompt loading. The loader passed .prompty markdown to gray-matter without disabling its executable 'js'/'javascript' frontmatter engines, so a '---js' frontmatter block is evaluated as Node.js at parse time. Exploitation requires the victim's application to load the malicious file (UI:P); no public exploit is identified at time of analysis and it is not on CISA KEV, but the fix commit ships a working proof-of-concept test.
OS command injection in Pheditor 2.0.1 through 2.0.5 lets an authenticated user with the default-enabled `terminal` permission bypass the TERMINAL_COMMANDS allowlist and run arbitrary shell commands as the web server user. The flaw is an incomplete fix for GHSA-9643-6xjp-vx57: the sanitization blocklist added `$` but still misses the single pipe `|`, backtick, and newline (0x0A), each of which starts with a whitelisted prefix and slips past both filters. A fully working exploit is published in the GHSA advisory, and because Pheditor ships with the default password `admin`, the required authentication is frequently trivial to obtain, effectively lowering the barrier to unauthenticated RCE.
Privilege escalation in the WPFunnels (Funnel Builder for WooCommerce) plugin for WordPress lets authenticated users holding the wpf_manage_funnels capability — typically the plugin's Funnel Manager custom role — rewrite the global wp_user_roles option and grant their own role full administrator capabilities. All versions up to and including 3.12.8 are affected, and Wordfence rates it CVSS 8.8 (High). No public exploit has been identified at time of analysis, but the root cause is trivial (an unvalidated option name reaching update_option()), making reliable weaponization straightforward for a low-privileged insider.
Privilege escalation in the Digits WordPress Mobile Number Signup and Login plugin (all versions through 9.1.0.5) lets authenticated Subscriber-level users promote themselves to Administrator by submitting a forged digits_reg_userrole value during a profile update. The flaw stems from the dig_update_wpwc_custom_fields() function failing to validate the caller's authorization or the requested role. There is no public exploit identified at time of analysis, but the low barrier (any registered user on an affected site with the DIGITS User Role field configured) and full-administrator outcome make this a high-priority patching item.
SQL injection in LogicalDOC Enterprise (versions up to and including 9.1.1) lets an authenticated user tamper with backend database queries through the ComparisonServlet component, which fails to properly sanitize user-supplied input. Because it is a blind injection, an attacker infers data through boolean/time-based inference rather than direct output, and the CVSS 8.8 vector (PR:L) indicates any low-privileged authenticated account can achieve high confidentiality, integrity, and availability impact. A public technical disclosure (netero1010) documents the flaw, but it is not listed in CISA KEV and EPSS is low (0.17%, 6th percentile), so widespread automated exploitation is not currently indicated.
Sandbox escape in PipeWire's PulseAudio compatibility layer lets a low-privileged process inside a confined environment such as Flatpak load an attacker-controlled library and execute arbitrary code on the host user's session. The flaw (CWE-427) is present in PipeWire as shipped in Red Hat Enterprise Linux 7, 8, 9 and 10 and carries a scope-changing CVSS 3.1 score of 8.8. There is no public exploit identified at time of analysis, and it is not listed in CISA KEV.
Divide-by-zero in AutomationDirect Productivity Suite enables a local, low-privileged attacker to crash the application, resulting in a denial-of-service condition on engineering workstations used to program AutomationDirect PLCs. All versions are affected per CPE wildcard data, and the vulnerability was disclosed via CISA ICS advisory ICSA-26-197-04, indicating ICS-CERT involvement and an OT/industrial context. No public exploit or CISA KEV listing has been identified at time of analysis, and exploitation impact is limited strictly to availability - no data exfiltration or tampering is possible.
Authenticated command-injection in Pheditor 2.0.4 lets any user holding the 'terminal' permission bypass the TERMINAL_COMMANDS allowlist and run arbitrary OS commands as the web server user. The terminal handler only blocks '&', ';', and '||' and validates commands with a prefix check before handing the full string to shell_exec(), so shell substitution like ls$(...) satisfies the allowlist while executing attacker-controlled code. A working PoC is published in the GitHub advisory, though there is no public exploit identified as a standalone weaponized tool and no active exploitation reported.
PowerShell command injection in Lenovo XClarity Integrator for Microsoft Windows Admin Center (versions 5.1.1 and below) running on the WAC Gateway allows an authenticated low-privileged attacker to inject arbitrary PowerShell commands when the plugin establishes remote PowerShell sessions. Successful exploitation yields high-impact code execution that extends beyond the plugin into subsequent systems (managed hosts), with full confidentiality, integrity, and availability compromise. No public exploit identified at time of analysis, and the issue is not listed in CISA KEV.
Remote code execution in the Loco Translate WordPress plugin (all versions up to and including 2.8.5) is reachable through a Cross-Site Request Forgery flaw in the execTemplate function, where missing nonce validation lets an unauthenticated attacker who tricks a logged-in administrator into clicking a crafted link supply a php://filter stream wrapper as the 'template' parameter, bypassing path validation and reaching a PHP include sink to execute arbitrary code. The chain converts a classic CSRF (CWE-352) into full server compromise, carrying CVSS 8.8 (C:H/I:H/A:H). There is no public exploit identified at time of analysis and it is not listed in CISA KEV.
Arbitrary file read in the Whistle debugging proxy (avwo/whistle) before 2.10.3 lets remote attackers retrieve any file on the host by abusing the GET /cgi-bin/temp/get endpoint. Because lib/service/service.js only sanitizes req.query.filename when it matches the temporary-file pattern and otherwise hands the raw filename to getFile, a value like ../../../../etc/passwd escapes TEMP_FILES_PATH and is served back. The CVSS 4.0 score is 8.7 (high) with a confidentiality-only impact; there is no public exploit identified at time of analysis, and the flaw is fixed in 2.10.3.
Denial of service in stoatchat (delta) versions before 20250210-1 (0.8.2) lets a remote unauthenticated attacker exhaust server resources by abusing the 'nearby' message query route. A logic error passes a message limit of zero to the database, which MongoDB interprets as 'no limit', so a single crafted request downloads an entire channel's history; parallelized requests amplify this into resource-exhaustion DoS. No public exploit identified at time of analysis and the flaw is not in CISA KEV, but the upstream fix and a VulnCheck advisory are published.
Uncontrolled memory and CPU exhaustion in Centrifugo real-time messaging server before 6.8.4 lets unauthenticated attackers send a small compressed WebSocket frame to the unidirectional transport that inflates into a massive decompressed payload. The message_size_limit was only checked against compressed wire bytes, while decompressed output was read via io.ReadAll with no cap, enabling a classic decompression bomb denial of service. No public exploit is identified at time of analysis, and the issue is not listed in CISA KEV.
SQL injection in the Quix Page Builder Pro extension for Joomla lets remote unauthenticated attackers inject arbitrary SQL through a crafted request, enabling database read and write against affected sites. The CVSS 4.0 base score is 8.7 (High), reflecting network vector, low complexity, no privileges, and high impact to confidentiality, integrity, and availability of the vulnerable database. There is no public exploit identified at time of analysis and the issue is not listed in CISA KEV; a third-party disclosure write-up exists on mysites.guru.
Missing authentication in Microsoft SimpleChat (prior to 0.241.206) lets remote, unauthenticated clients invoke administrative plugin validation, health-check, and repair endpoints that were documented as protected but never enforced authentication at runtime. Any network-reachable attacker can trigger plugin instantiation, health probing, and repair actions without login or admin rights. No public exploit identified at time of analysis, and the CVE is not listed in CISA KEV.
Reflected cross-site scripting in Emlog 2.6.13 and earlier lets a remote attacker execute arbitrary JavaScript in an authenticated administrator's backend session by luring the admin to open a crafted link that hits the user-search feature in admin/user.php. The keyword parameter is sanitized with addslashes but never HTML-escaped before it is echoed into a value attribute in admin/views/user.php, so injected markup breaks out of the attribute and runs in the admin's browser. No public exploit identified at time of analysis, and no fixed version currently exists.
Local privilege escalation in ESET Inspect Connector for Windows lets a low-privileged local user escalate to elevated (SYSTEM-level) privileges by abusing an improperly authenticated ALPC (Asynchronous Local Procedure Call) IPC channel. The affected component is the endpoint agent that links ESET Inspect (EDR) to protected Windows hosts; a non-privileged process can issue unauthenticated requests to a privileged service endpoint that fails to verify the caller. The issue was reported and fixed by ESET, is not listed in CISA KEV, and has no public exploit identified at time of analysis.
Brute-force protection bypass in Perforce Delphix Continuous Data lets remote attackers defeat the account lockout mechanism by firing concurrent authentication requests that race the failed-login counter, allowing unlimited password guessing against a targeted account. The flaw is a time-of-check/time-of-use race (CWE-307) rather than a direct code-execution bug, and no public exploit identified at time of analysis. It matters because it silently neutralizes a control administrators rely on to stop credential-stuffing and password-spraying.
Authentication bypass in HCL DFXServer allows remote unauthenticated attackers to reach specific API endpoints directly and perform unauthorized actions without valid credentials, because the application never verifies authentication status on those routes. The primary impact is high confidentiality exposure with limited integrity impact (CVSS 8.2), and the flaw was self-reported by HCL. There is no public exploit identified at time of analysis and it is not listed in CISA KEV.
Authentication bypass in HCL DFXServer lets an unauthenticated remote attacker intercept and tamper with the server's authentication responses to be granted access as a legitimate user without presenting valid credentials. Rated CVSS 8.2 (high) with CWE-294 (Authentication Bypass by Capture-replay), the flaw exposes confidential application data to any network-reachable attacker. No public exploit identified at time of analysis and it is not listed in CISA KEV.
Arbitrary file deletion in the Uncanny Automator WordPress plugin (all versions through 7.3.1.4) lets unauthenticated attackers delete any file on the server via untrusted PHP object deserialization in the fr_token function, and deleting a critical file such as wp-config.php can pivute WordPress into an installation/setup state that yields remote code execution. Exploitation is gated by a specific setup: a Forminator form must be connected to an Uncanny Automator recipe whose trigger is configured for 'Everyone', which is what exposes the deserialization sink to anonymous form submissions. A self-contained gadget chain via Action_Helpers_Email::__destruct() ships inside the plugin, so no external gadget library is needed; reported by Wordfence with no public exploit identified at time of analysis and no CISA KEV listing.
Access-control bypass in Keycloak (fixed in 26.5.3) lets a disabled user account still obtain valid tokens through the JWT authorization grant preview feature. When that feature is enabled, Keycloak omits the account-enabled check during JWT authorization grant processing, so a low-privileged remote attacker who can present a valid assertion token from an external identity provider can mint a JWT for an account an administrator has explicitly disabled, regaining access to protected resources. No public exploit identified at time of analysis and it is not listed in CISA KEV.
Authentication/access-control bypass in BigBlueButton's bbb-web component (versions prior to 3.0.21) lets remote attackers reach protected API endpoints without a valid checksum by supplying the presentationUploadExternalUrl parameter. BigBlueButton's API is normally guarded by a shared-secret checksum, so bypassing it undermines the core authorization mechanism for calls such as CreateMeeting. No public exploit identified at time of analysis, though the fixing commit is public; CVSS is 8.1 and there is no CISA KEV listing or EPSS signal in the provided data.
Session hijacking in BigBlueButton before 3.0.21 lets an authenticated conference participant predict and reuse other users' session tokens to impersonate them. The bbb-web component generated 16-character sessionToken values (and 12-character auth tokens and internal user IDs) using Apache Commons RandomStringUtils.randomAlphanumeric, a non-cryptographic PRNG, making the token space guessable to anyone who already holds a valid session. Rated CVSS 8.1 with high confidentiality and integrity impact; no public exploit identified at time of analysis.
Missing-authorization schema tampering in ArcadeDB before 26.6.1 lets any authenticated identity - including a read-only API token that lacks the UPDATE_SCHEMA permission - mutate the database schema over the HTTP command/query endpoints. Because CVE-2026-44221's fix only added the UPDATE_SCHEMA check to LocalDocumentType.createProperty, the remaining schema mutators (DROP PROPERTY, ALTER TYPE, ALTER PROPERTY, type/bucket/alias changes) stayed unguarded, so a low-privilege user can drop properties, rename or re-parent types, and flip property constraints, corrupting the meaning of every stored record. There is no public exploit identified at time of analysis and it is not in CISA KEV, but the vendor advisory (GHSA-vg6x-6pg9-6qwg) confirms the flaw and a fix in 26.6.1.
Information disclosure and memory-safety defect in the Linux kernel's IPv4 output path (__ip_append_data) affects kernels from 6.0 up to the fixed 6.6.144, 6.12.95, 6.18.38 and 7.1.3 releases, where the paged-allocation branch under-sizes the new skb's linear area by the fraggap byte count while overstating pagedlen. A local user driving corked/fragmented socket writes can trigger the miscalculation, potentially leaking adjacent kernel memory (vendor tags classify it as Information Disclosure) or causing corruption. There is no public exploit identified at time of analysis, EPSS is low at 0.19% (9th percentile), and it is not listed in CISA KEV.
Heap use-after-free in YAML::Syck for Perl before version 1.47 lets a crafted YAML document corrupt parser memory when it redefines a previously declared anchor. The bug lives in the bundled libsyck C code (syck_hdlr_add_anchor), where an anchor name string is shared between a node's node->anchor field and the parser's anchors-table key; freeing the node frees the shared key, and a subsequent anchor redefinition makes st_delete/st_strcmp compare against freed heap memory. Any Perl application that calls Load or LoadFile on untrusted YAML is affected on the default code path (anchors need no special flags). Reported by CPANSec and disclosed via oss-security; publicly documented proof-of-concept trigger inputs exist, but there is no CISA KEV listing and no evidence of active exploitation.
Local privilege escalation in Zoom Rooms for Windows before version 7.1.0 allows an authenticated low-privileged user to elevate to higher privileges via local access, with high impact to confidentiality, integrity, and availability (CVSS 7.8). Zoom's own security team (security@zoom.us) reported the flaw, classified as improper privilege management. No public exploit identified at time of analysis, and it is not listed in CISA KEV.
Out-of-bounds heap write in QEMU lets a local attacker operating inside a guest virtual machine corrupt memory in the host emulator process, enabling information disclosure, data-integrity corruption, or denial of service. The flaw stems from cpu_physical_memory_map() returning a shorter buffer length than the caller assumes, so subsequent writes spill past the allocation. Rated CVSS 7.8 (CWE-787); no public exploit identified at time of analysis, but multiple distributions (Red Hat, Ubuntu, SUSE, Debian) have shipped fixes.
Local privilege escalation and silent denial-of-service in the Wazuh Windows agent (versions 4.6.0 through 4.14.4) arise from a heap-based buffer overflow in the syscheck registry-monitoring component. Because wazuh-agent.exe runs as NT AUTHORITY\SYSTEM, a low-privileged local user who creates a 255-character registry subkey under a monitored wildcard path can trigger an out-of-bounds heap write, either blinding the agent or potentially executing code as SYSTEM. Proof-of-concept exploit code exists (SSVC: poc), but it is not in CISA KEV and EPSS is very low (0.09%, 1st percentile), consistent with the local-only, non-automatable attack profile.
Code injection in Microsoft Kiota Ruby code generator allows remote code execution through crafted OpenAPI specifications. Kiota versions prior to 1.32.0 fail to escape '#' characters in schema-derived strings embedded in Ruby double-quoted literals, enabling Ruby interpolation sequences such as '#{code}' to execute arbitrary commands when the generated client code runs. While no active exploitation is confirmed, the attack is automatable, EPSS indicates moderate probability (1.47%), and the vendor has released a fix.
Local file inclusion in Emlog 2.6.13 and earlier lets an authenticated author supply a path-traversal template value that is stored unvalidated by api_controller.php and later passed to View::getView() in log_controller.php, causing an arbitrary local .php file to be included when the article is viewed. Because included PHP is executed, an attacker who can place or influence a .php file on the host can escalate this to code execution. No public exploit has been identified at time of analysis, and no fixed version currently exists.
Out-of-bounds heap read in YAML::Syck before 1.47 for Perl lets an attacker who controls a parsed YAML document trigger a one-byte over-read in the bundled libsyck lexer, with a total (per SSVC) technical impact of information disclosure or process crash. The flaw sits in newline_len/is_newline, which dereference the scan pointer (and the following byte for a "\r\n" pair) with no NUL-terminator or bounds check, so a block scalar sitting at a document boundary runs one byte past the heap lexer buffer. There is no public exploit identified at time of analysis and it is not in CISA KEV, but CPANSec-documented trigger inputs exist and it is explicitly an incomplete fix of CVE-2025-11683 on a lexer path the earlier patch missed.
We need to produce a comprehensive analysis JSON for CVE-2026-53598, based on the provided data. The CVE involves Microsoft's Prompty (a markdown file format for LLM prompts), path traversal in file references, fixed in 2.0.0-beta.2. We have CVSS: 7.5, vector AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:N/A:N, CWE-22. EPSS 1.06% (percentile 61%). Patch available. SSVC: Exploitation none, Automatable yes, Technical impact partial. CPE: cpe:2.3:a:microsoft:prompty:*:*:*:*:*:*:*:*. Versions fixed: 2.0.0b2 for PyPI, 2.0.0-beta.2 for npm, crates.io, NuGet. We have a GitHub Advisory detailing the vulnerability, affected packages, fix details. The commit diff shows the fix canonicalizing paths, restricting file references to within the prompt directory by default, and allowing host apps to provide additional allowed roots. The diff adds tests for traversal, absolute path, and symlink escape. Now we need to fill in the JSON fields. We'll synthesize from all sources. product_name: "Prompty" - the specific product name, as per instructions (1-3 words). "Microsoft Prompty" could work, but "Prompty" is sufficient. I'll use "Microsoft Prompty" to include vendor, but requirement: "NOT the vendor alone, NOT a description." So "Prompty" alone is fine and concise. Using "Prompty". summary: We must not copy or paraphrase description. Original synthesis using all data. Start with specific impact verb and product. So e.g., "Arbitrary file read in Microsoft Prompty before 2.0.0-beta.2 allows an attacker to read local files via crafted .prompty frontmatter file references." Mention that it allows path traversal, absolute paths, symlink escapes. Not actively exploited (KEV false, EPSS 1.06% low), POC exists? Not mentioned, but there are regression tests that demonstrate the attack, which could serve as proof-of-concept. But we'll state: "no public exploit identified at time of analysis" per rules, because no POC tag is provided. Wait, there's no explicit POC=YES in input, only the advisory and commi
Code injection in Microsoft Kiota's Python client generator (versions prior to 1.32.0) allows an attacker who controls an OpenAPI description to embed arbitrary Python that executes when a developer generates and imports the resulting client. The flaw stems from unsanitized x-ms-enum.values[].description strings being emitted as inline comments; embedded newlines break out of the comment and run at module scope. No public exploit identified at time of analysis, and the issue is not in CISA KEV. The assigned CVSS 3.1 base score is 7.5, though its impact metrics (C:N/I:N/A:H) understate what is functionally full code execution.
Server-side request forgery in Stoatchat before 0.14.0 lets unauthenticated, network-accessible attackers reach internal services that a DNS-based IP blocklist was meant to protect. The url_is_blacklisted function validates only the first resolved address, while the underlying HTTP client iterates over all cached DNS addresses, so a hostname resolving to both an allowed and an internal/blocked IP slips past the check. No public exploit identified at time of analysis; the issue is fixed in the v0.14.0 release and was reported by VulnCheck.
Denial of service in libsolv's PGP signature verification (solv_pgpvrfy) allows remote attackers to crash automated package and repository processing by supplying a crafted Ed25519 signature. The flaw is a stack-based buffer overflow (CWE-121) triggered when the EdDSA 's' MPI is copied into a fixed 64-byte stack buffer using an attacker-controlled mismatched length. It affects the libsolv dependency solver embedded in Red Hat Enterprise Linux 7/8/9/10, OpenShift, Satellite, RHUI, openSUSE/SUSE, Ubuntu and Debian; no public exploit identified at time of analysis and it is not in CISA KEV.
Webhook token disclosure in stoatchat (the Rust 'delta' backend of the Revolt/stoatchat platform) lets any channel member holding only ViewChannel (read) permission enumerate a channel's webhooks and read their secret tokens. Because a retrieved token authorizes posting, the attacker can then inject arbitrary messages while impersonating a bot or webhook, bypassing the channel's ManageWebhooks and posting permissions. Reported by VulnCheck and fixed in 20250210-1 (0.8.2); no public exploit identified at time of analysis.
Denial of service in the Feast Feature Server (the Python feature-store serving component, also shipped within Red Hat OpenShift AI/RHOAI) lets remote unauthenticated attackers exhaust host resources through its `/ws/chat` WebSocket endpoint. Because the endpoint accepts unlimited concurrent connections and unbounded message traffic, opening many simultaneous sockets drains memory, CPU, and file descriptors until legitimate clients are locked out. There is no public exploit identified at time of analysis and this is not listed in CISA KEV; the upstream fix in PR #192 caps connections, message size, and message rate.
Memory-exhaustion denial of service in Meta's Proxygen HTTP library (versions v2017.01.16.00 through v2026.07.20.00) allows remote, unauthenticated attackers to force unbounded server-side buffering by abusing HTTP/2 flow control. Because the core HTTP session layer had no generalized slow-consumer detection, a client that advertises SETTINGS_INITIAL_WINDOW_SIZE of 0 or simply withholds WINDOW_UPDATE frames can stall many concurrent streams while fully-generated response bodies pile up in memory. No public exploit identified at time of analysis, but the technique is well-understood and the CVSS 7.5 availability impact is real against default deployments.
Remote code execution in Microsoft Windows Terminal allows an unauthenticated attacker to run arbitrary code by tricking a user into interacting with malicious content that triggers an integer overflow (CWE-190). The flaw carries a CVSS 7.5 rating driven by high complexity and required user interaction; no public exploit has been identified at time of analysis, and it is not listed in CISA KEV. Microsoft has released a patch through the MSRC update guide.
Remote denial of service in CoreDNS versions prior to 1.14.4 allows an unauthenticated attacker to crash the DNS server with a single 28-byte UDP datagram when the proxyproto plugin is enabled. The malformed PROXY protocol v2 header triggers a nil pointer dereference in the PacketConn.ReadFrom log path, killing the process before per-request recovery can intervene. No public exploit identified at time of analysis, though the fixing PR includes a reproducer test that encodes the exact crashing byte sequence; not listed in CISA KEV.
SQL injection in ThemeHunk's Advance Product Search - Voice & Ajax Search for WooCommerce plugin (all versions ≤ 1.4.4) lets unauthenticated attackers inject arbitrary SQL through the 's' and 'match' search parameters, enabling extraction of sensitive database contents such as user credentials and order data. The flaw stems from unescaped, unprepared query construction in the plugin's AJAX search backend and is remotely exploitable without authentication or user interaction. Reported by Wordfence; no public exploit identified at time of analysis and not listed in CISA KEV.
Denial of service in the h2o HTTP server (all versions prior to commit 6b5370d) allows remote unauthenticated attackers to crash the server by requesting static files whose on-stack path construction via alloca can reach ~600KB, exceeding musl libc's default 128KB pthread stack and triggering a guard-page segmentation fault. The flaw is a CWE-770 uncontrolled resource allocation issue affecting availability only (CVSS 7.5, A:H). There is no public exploit identified at time of analysis and it is not listed in CISA KEV, but exploitation is trivial where h2o is built against musl libc.
Denial of service in Quicly, the QUIC protocol library embedded in the H2O HTTP server, allows remote attackers to corrupt connection state and crash the process prior to commit 8b178e6. Quicly's packet decoder accepts Connection IDs up to 255 bytes (permitted for unknown QUIC versions) but its version-1 CID buffers are only 20 bytes, and the library - unlike its bundled CLI - never rejected over-length CIDs, letting an oversized CID overrun the buffer within its allocated chunk and trigger assertion failures. No public exploit is identified at time of analysis, and the flaw is not listed in CISA KEV.
Remote denial of service in quicly, the IETF QUIC implementation used by the H2O HTTP server, allows unauthenticated attackers to crash the process by sending more than 32KB of valid handshake messages over the CRYPTO stream within a single packet number space, tripping a reachable assertion (CWE-617). Affects all builds prior to commit 937d0e9. No public exploit identified at time of analysis and the flaw is not in CISA KEV, but the network-reachable, no-privilege nature makes it trivially triggerable once understood.
Uncontrolled memory allocation in Grafana Loki lets remote unauthenticated users issue queries with excessively large limit values, forcing the service to allocate large amounts of memory and potentially exhausting host resources. The impact is denial of service (CVSS 7.5, A:H only) and varies by deployment topology - a single-binary or under-provisioned deployment is far more exposed than a horizontally scaled microservices setup. There is no public exploit identified at time of analysis and it is not listed in CISA KEV; note the input 'Information Disclosure' tag conflicts with the CVSS impact (C:N/I:N/A:H), which indicates an availability-only issue.
Memory exhaustion in Quicly's QUIC flow control handling allows an unauthenticated remote peer to force disproportionate receive buffer allocation in the H2O HTTP server by sending STREAM frames containing a single byte placed at the maximum permitted stream offset. The library pre-allocates buffers up to the largest received offset to accommodate out-of-order delivery, so a handful of crafted packets can cause the server to allocate memory far exceeding what the transferred data warrants. Under H2O's default stream concurrency settings, this attack multiplies per-connection memory usage by roughly 4x, making memory exhaustion achievable at low bandwidth cost. No public exploit identified at time of analysis; EPSS is 0.25% (16th percentile) and CISA SSVC lists exploitation as 'none', suggesting no observed real-world abuse.