Remote code execution in GitLab CE/EE package registry exposes instances across three active version branches to full host compromise by an authenticated low-privilege user. Path traversal (CWE-22) in the package registry component allows directory escape and arbitrary file write under certain server-side conditions, with CVSS scope change (S:C) confirming that impact extends beyond the GitLab application to the underlying host. A publicly available exploit exists via HackerOne report 3754194, though this vulnerability is not listed in the CISA KEV catalog at time of analysis.
Uncontrolled recursion in the justhtml Python HTML parsing library (all versions through 1.9.1) allows any remote unauthenticated attacker who can supply HTML input to trigger a denial of service via stack exhaustion. Approximately 1,000 nested HTML elements (~11 KB of crafted input) are sufficient to exceed CPython's default recursion limit of 1,000 frames during DOM traversal in TreeBuilder._find_elements(), raising an unhandled RecursionError that can abort parsing, fail requests, or terminate a worker process depending on framework error handling. Vendor patch 1.10.0 is available; no public exploit code or CISA KEV listing has been identified at time of analysis, though the advisory's code-level detail makes independent reproduction trivial.
Privilege escalation in the OnGres StackGres operator lets a low-privilege, database-owning tenant elevate to administrator privileges, breaking the multi-tenant isolation model of this Kubernetes-based PostgreSQL platform. The CVSS 9.9 rating with a scope change (S:C) reflects that an authenticated tenant confined to their own database can cross a trust boundary to gain full administrative control. No public exploit identified at time of analysis, and the issue is not listed in CISA KEV; the only public reference is a GitLab work item tracking the flaw.
Heap out-of-bounds write in DBD::Pg 3.21.0, the Perl DBI driver for PostgreSQL, lets an attacker corrupt heap memory by passing special floating-point literals (NaN, Inf, +Inf, -Inf, Infinity, and signed variants) to the quote_float() routine, reachable through the common $dbh->quote(value, DBI::SQL_NUMERIC) path. The bug is a 2-byte overwrite (a single quote plus a NULL terminator) introduced by the 3.21.0 quote.c rewrite and fixed in 3.21.1. No public exploit identified at time of analysis; EPSS is low (0.23%, 14th percentile) and CISA SSVC records exploitation status as none.
Unauthenticated privilege escalation to administrator in the RestrictMate WordPress plugin before 1.3.0 lets any remote visitor abuse the account-registration flow, which fails to validate the user role submitted during signup, to create a brand-new administrator account and receive a logged-in admin session - resulting in full site takeover. Rated CVSS 9.8 with a fully remote, no-interaction, no-privilege vector; a vendor patch (1.3.0) exists and CISA's SSVC framework flags it as automatable with total technical impact. There is no public exploit identified at time of analysis and it is not on CISA KEV, and EPSS is low at 0.19% (8th percentile).
SQL injection in code-projects Barangay Resident Profiling Management System 1.0 allows unauthenticated remote attackers to manipulate database queries via the Search argument in residents.php. The Resident Search Functionality fails to sanitize user-supplied input before constructing SQL statements, enabling read, write, or denial-of-service impact against the underlying database. A public proof-of-concept exploit is available on GitHub, lowering the bar for opportunistic exploitation; however, the product's extremely limited real-world deployment significantly constrains actual attack surface.
Sanitizer bypass in the justhtml Python library (versions <= 1.11.0) allows attacker-controlled input to survive as raw HTML in the output of to_markdown(), enabling cross-site scripting when that Markdown is later rendered. Although to_html() safely neutralizes the same input, the Markdown serializer escapes only a handful of Markdown metacharacters and leaves angle brackets (< and >) intact, so entity-decoded text (e.g. <script>) or text from RCDATA/RAWTEXT elements like <title>, <textarea>, <noscript>, and <plaintext> is emitted verbatim. Publicly available exploit code exists (a proof-of-concept is published in the vendor advisory), and the flaw is fixed in 1.12.0.
Cross-site scripting via HTML sanitization bypass affects justhtml (a Python HTML parsing/sanitization library by EmilStenstrom) in all versions up to and including 1.15.0, letting active content such as script or style survive sanitization. The flaws mostly hit advanced usage - programmatic DOM input, reused or mutated policy objects, and custom policies that preserve SVG/MathML - rather than the default JustHTML(..., sanitize=True) path, though process-wide mutation of exported default policies can weaken sanitization broadly. Reported by VulnCheck and fixed in 1.16.0; no public exploit identified at time of analysis and it is not listed in CISA KEV.
Sanitization bypass in justhtml (a Python HTML/Markdown sanitization library) before 1.15.0 lets attacker-controlled input reintroduce active HTML and JavaScript, resulting in cross-site scripting in consuming web applications. The flaws span URL-sanitization helpers, HTML serialization, Markdown passthrough, and custom policies - most do NOT affect the default sanitize=True path but do affect helper APIs, programmatic DOM construction, html_passthrough=True, and custom policies. No public exploit identified at time of analysis; the issue was reported by VulnCheck and fixed in 1.15.0.
PHP Object Injection in PPWP - Password Protect Pages (WordPress plugin, versions up to and including 1.9.18) permits authenticated attackers holding Contributor-level roles or higher to deserialize attacker-controlled data via the 'post_protection_roles' parameter, injecting arbitrary PHP objects into the application. Real-world impact is fully contingent on a Property-Oriented Programming (POP) chain being present in another co-installed plugin or theme - absent that chain, the vulnerability has no exploitable effect, but where one exists, outcomes can include arbitrary file deletion, sensitive data retrieval, or remote code execution. No public exploit code or active exploitation has been identified at time of analysis; however, a Wordfence advisory and a WordPress SVN changeset (3567221) both indicate a fix has been committed.
Privilege escalation in the Security Hardener WordPress plugin (all versions up to and including 2.4.4) allows any authenticated Subscriber-level user to create new Administrator accounts or reset existing Administrator passwords by sending crafted REST API requests. The plugin's user-enumeration blocking feature inadvertently replaces WordPress Core's granular capability checks on the /wp/v2/users REST endpoints with a bare is_user_logged_in() check, completely stripping create_users, edit_users, and delete_users enforcement. Because block_user_enum is enabled by default, this authorization bypass is active on every site immediately after plugin installation with no additional configuration required; no public exploit has been identified at time of analysis.
Uncontrolled resource consumption in justhtml before 1.18.0 enables denial-of-service conditions through CPU and memory exhaustion when processing attacker-controlled CSS selectors, large untrusted documents, cyclic DOM graphs, or linkification input. The impact is strictly availability - the advisory explicitly confirms no script execution, data disclosure, or sanitizer bypass is possible. No public exploit has been identified at time of analysis, and default JustHTML(sanitize=True) configurations are not expected to be exposed; risk is confined to applications that forward untrusted input into selector-evaluation or linkification APIs.
Eval injection in chirpmyradio CHIRP's Kenwood ITM driver allows arbitrary Python code execution when a user opens a specially crafted CSV file. The vulnerable `_clean_tmode` function in `drivers/kenwood_itm.py` passes untrusted TXSIG and RXSIG CSV column values directly to Python's built-in `eval()` without sanitization, confirmed by the upstream commit diff. A proof-of-concept repository exists publicly, though no active exploitation is confirmed in CISA KEV. All CHIRP builds prior to commit 39178dbfc4fece083ab9ed20286d6ae3a91a718e are affected.
Reflected cross-site scripting in the Brave Popup Builder WordPress plugin (versions up to 0.8.5) allows unauthenticated remote attackers to inject arbitrary JavaScript into a victim's browser by crafting a malicious URL containing unsanitized UTM query parameters that are reflected directly into popup form HTML. Exploitation requires the victim to follow a crafted link, making phishing the primary delivery mechanism. No public exploit code or CISA KEV listing has been identified at time of analysis, and the low EPSS score (0.21%, 11th percentile) suggests limited real-world exploitation activity; nonetheless, the fix is available and upgrade is straightforward.
Null pointer dereference in llama.cpp's ggml-RPC server (commit bec4772f6) allows remote unauthenticated denial of service against any deployment exposing the RPC port to the network. The flaw resides in rpc_server::graph_compute, where a graph node with id=0 bypasses a conditional null check, enabling a single crafted TCP packet to crash the server process. No standalone exploit code is publicly available at time of analysis, though the upstream fix PR includes a Python test script that reproducibly triggers the crash in under one second.
Improper input validation in llama.cpp's ggml-RPC Server allows remote unauthenticated attackers to send crafted deserialization payloads targeting the `deserialize_tensor` function, affecting the `op` and `op_params` arguments in `ggml/src/ggml-rpc/ggml-rpc.cpp` at commit bec4772f6. The CVSS 4.0 score of 6.9 (AV:N/AC:L/PR:N/UI:N) reflects network-accessible exploitation with partial confidentiality, integrity, and availability impact on the vulnerable system. No public exploit code or active exploitation (CISA KEV) has been identified at time of analysis, and the originating GitHub issue was closed due to inactivity.
Directory traversal in WebToffee WooCommerce PDF Invoices, Packing Slips, Delivery Notes & Shipping Labels plugin (all versions ≤ 4.9.8) allows subscriber-level authenticated attackers to read arbitrary server files via the get_image_src_in_base64 function. The exfiltration channel is particularly stealthy: base64-encoded file contents are embedded into cached invoice HTML and retrieved through the plugin's own legitimate Print/Download endpoints using only a valid nonce and access key, potentially bypassing WAF rules that inspect request paths. No public exploit or CISA KEV listing is identified at time of analysis; however, the trivially low privilege bar - any registered WooCommerce customer qualifies - combined with high-value targets like wp-config.php credentials makes this an operational priority for any public-facing WooCommerce store.
Command injection in Tenda CH22 version 1.0.0.1 allows a remote low-privileged attacker to execute arbitrary OS commands by manipulating the `editNameMit` parameter within the `formeditFileName` function at the `/goform/editFileName` web management endpoint. A public proof-of-concept exploit has been released, confirmed by the CVSS 4.0 E:P modifier and a linked Notion-hosted exploit writeup. No CISA KEV listing at time of analysis, but the public PoC significantly lowers the skill bar for exploitation against any CH22 device with an exposed or accessible management interface.
Command injection in Tenda CH22 firmware version 1.0.0.1 allows a low-privileged remote attacker to execute arbitrary OS-level commands via the cmdinput parameter of the /goform/exeCommand endpoint. The underlying formexeCommand function passes user-controlled input to a system command without adequate sanitization, exposing the router's underlying Linux OS. A proof-of-concept exploit has been publicly disclosed on Notion (referenced via VulDB), elevating real-world risk beyond the moderate CVSS 4.0 score of 5.3, which appears to understate the practical OS-level impact typical of router command injection primitives.
Cross-site scripting in SourceCodester Stock Management System 1.0 allows remote unauthenticated attackers to inject arbitrary JavaScript via the clientName and clientContact parameters of /php_action/getOrderReport.php. A publicly available proof-of-concept exploit exists (CVSS 4.0 E:P), lowering the skill threshold for exploitation. No CISA KEV listing is present, but the unauthenticated, network-accessible vector with a public PoC makes this a credible risk for any exposed deployment.
Cross-site scripting in SourceCodester Stock Management System 1.0 allows remote attackers to inject malicious script via the clientName and clientContact parameters of /php_action/printOrder.php. When a legitimate user views a manipulated order print page, the unsanitized input is rendered in their browser, enabling session hijacking, credential theft, or malicious redirects. A public exploit has been disclosed via GitHub (referenced in VulDB entry 394299); no active exploitation is confirmed in CISA KEV.
Cross-site scripting in SourceCodester Class and Exam Timetabling System 1.0 allows remote attackers to inject arbitrary JavaScript via the unsanitized `course` parameter in `/BSIT2.php`. Exploitation requires victim interaction (UI:P per CVSS 4.0) but no authentication (PR:N), enabling session hijacking or phishing against users of the affected application. A public proof-of-concept exploit has been disclosed on GitHub; no CISA KEV listing indicates no confirmed widespread active exploitation at time of analysis.
Reflected cross-site scripting in SourceCodester Class and Exam Timetabling System 1.0 allows remote attackers to inject and execute arbitrary JavaScript in a victim's browser by manipulating the `course` parameter of /BSIS1.php. A publicly available proof-of-concept exploit exists (referenced via GitHub), raising the practical threat level above what the moderate CVSS 4.0 score of 5.3 alone would suggest. No authentication is required to craft the malicious request, though successful exploitation depends on user interaction - a victim must visit a attacker-controlled or attacker-supplied URL.
Open redirect via safe-URL validator bypass in CTFd versions 3.8.0 through 3.8.3 allows remote unauthenticated attackers to craft login URLs that redirect victims to attacker-controlled sites, enabling phishing of competition participants. The vulnerability lies in the _is_safe_url() function within CTFd/utils/validators/__init__.py, which fails to correctly reject malformed or protocol-relative values supplied via the 'Next' query parameter. The vendor CHANGELOG for the patched release (3.8.4) further notes that attacker-controlled links could serve as a vector for arbitrary JavaScript execution if clicked, suggesting the integrity impact may exceed the open-redirect framing; a public exploit writeup exists and no active KEV listing is present at time of analysis.
Improper authorization in SourceCodester Class and Exam Timetabling System 1.0 allows a low-privileged authenticated attacker to manipulate the `id` or `username` parameters in `/admin/edit_user_account.php` to modify user accounts they do not own or administer. The flaw is rooted in missing access control checks on the User Account Update endpoint, enabling horizontal or vertical privilege escalation by forging parameter values. A public proof-of-concept exploit has been disclosed on GitHub, raising the practical exploitation risk above what the moderate CVSS score alone suggests.
SQL injection in itsourcecode Hospital Management System Project in PHP 1.0 allows low-privilege authenticated remote attackers to manipulate database queries via the `delid` parameter in `/viewservicetype.php`. Publicly available exploit code exists, reported and disclosed through VulDB. Successful exploitation yields partial confidentiality, integrity, and availability impact against the application database, with no lateral scope change to external systems confirmed.
SQL injection in sambitraj's Student-Management-System exposes multiple user-supplied parameters in the Management Mutation Handler to database query manipulation by authenticated remote attackers. The application uses a rolling release model with no versioned releases; all commits up to 56ba287f2e9031523ccb4244cb6e3fe530e4e5d5 are affected and no patch has been issued despite responsible disclosure via GitHub issue. Publicly available exploit code exists (E:P per CVSS 4.0 supplemental), though no active exploitation has been confirmed and the CVSS 4.0 score of 2.1 reflects limited rated impact under low-privilege conditions.
SQL injection in the Dashboard component of sambitraj's Student-Management-System exposes the application to remote database manipulation via the roll_no and teacher_name parameters, affecting all commits through 56ba287f2e9031523ccb4244cb6e3fe530e4e5d5. A low-privileged authenticated attacker can inject arbitrary SQL to read or modify student records without any patch or vendor response currently available. A public exploit exists via GitHub issue disclosure, though real-world impact is constrained by the project's extremely narrow deployment footprint.
Authorization bypass via Insecure Direct Object Reference (IDOR) in the Boarder Management Module of code-projects Barangay Resident Profiling Management System 1.0 allows authenticated remote attackers to access or modify boarder records belonging to other users by manipulating the `ID` parameter in `/boarders.php`. A publicly available proof-of-concept exploit exists on GitHub, materially lowering the exploitation barrier. The vulnerability is not listed in CISA KEV, so no confirmed active exploitation campaigns have been identified at time of analysis.
Authorization bypass in code-projects Barangay Resident Profiling Management System 1.0 exposes the /archived_records.php endpoint to insecure direct object reference (IDOR) exploitation, allowing low-privileged authenticated attackers to restore or delete archived resident records they do not own by manipulating the resident_id parameter. A public proof-of-concept exploit is available via GitHub, though the vulnerability is not listed in CISA's Known Exploited Vulnerabilities catalog. The CVSS 4.0 score of 2.1 reflects limited confidentiality, integrity, and availability impact scoped to the vulnerable system with no lateral scope change.
Server-side template injection (SSTI) via the FreeMarker engine in Dromara UJCMS 10.1.0 through 10.1.3 allows authenticated backend administrators to inject malicious template directives through the web-file-template update API endpoint. Exploitation requires high-privilege backend credentials but is network-reachable with no user interaction, and a public proof-of-concept exploit has been published on GitHub demonstrating the attack path. The vendor-scored CVSS 4.0 impact is rated low across confidentiality, integrity, and availability, though the FreeMarker SSTI class is capable of full JVM-level code execution if template sandboxing is absent or misconfigured - defenders should not assume the impact ceiling matches the reported score.
Unauthenticated remote access to the cookies.txt file in MeTube (versions up to 2026.06.10) exposes yt-dlp session cookies via a direct HTTP request to the predictable path /download/.metube/cookies.txt. Any remote actor who can reach the MeTube instance can retrieve cookies potentially authenticating to YouTube or other streaming platforms, enabling session hijacking on those services. A public exploit has been disclosed; upgrading to version 2026.06.20 resolves the issue.
Authentication bypass in vas3k TaxHacker through version 0.8.2 is possible because the JWT Secret Handler in `lib/config.ts` falls back to a hard-coded value for `BETTER_AUTH_SECRET` when the environment variable is not explicitly set at deployment. Because TaxHacker is open-source on GitHub, the default secret is publicly readable, enabling any remote unauthenticated attacker to forge valid JWT tokens and impersonate arbitrary users - including administrators - against any misconfigured instance. No vendor patch exists; the maintainer had not responded to the disclosure issue at time of analysis, and publicly available exploit code is indicated by the CVSS 4.0 E:P flag.
Server-side request forgery in vas3k TaxHacker up to version 0.8.2 enables authenticated remote attackers to force the application server to initiate arbitrary IMAP TCP connections to attacker-controlled hosts and ports via the Email Sync configuration interface. The `buildImapConfig` function in `lib/email-sync/imap-client.ts` accepted unsanitized `host` and `port` parameters with no allowlist validation, no private-IP blocking, and no port restriction, allowing the TaxHacker server process to be redirected as a proxy into the internal network. No public exploit is identified at time of analysis; the upstream fix (PR #170) introduces a dedicated validation module but awaits merge acceptance and has not been released in a tagged version.
HTTP request smuggling in the Perl CPAN module Reverse::Proxy (versions before 0.04) allows unauthenticated remote attackers to inject arbitrary HTTP requests to upstream servers by embedding CRLF sequences in percent-encoded URL paths. The PSGI specification requires servers to percent-decode PATH_INFO before handing it to the application, meaning %0d%0a has already become a literal carriage return and line feed by the time the proxy appends it to the upstream request line - terminating that line early and promoting everything that follows to a second, attacker-controlled request. The injected request is attributed to the proxy by the upstream and can reach internal paths the proxy's own routing never exposes; on pooled keep-alive connections it may be delivered in the context of another client's session. No public exploit code and no CISA KEV listing exist at time of analysis.
Cross-site scripting in justhtml 0.9.0-1.21.0 allows an attacker who can control the text content of HTML code or pre elements to escape the Markdown inline code span produced by to_markdown(), causing sanitized content to be emitted unescaped and re-parsed as live HTML by a downstream Markdown renderer. The to_markdown() function is explicitly documented as a security boundary, making this a trust-model violation. No public exploit code has been identified and CISA KEV status is not confirmed, but the root cause is a fundamental gap in the inline code-span helper that bypasses the library's own stated sanitization guarantees.
Cross-site scripting in justhtml before 3.11.0 allows remote unauthenticated attackers to execute arbitrary JavaScript in victims' browsers via SVG or MathML elements injected through selectedcontent projections. The default sanitizer fails to remove event handlers that are cloned and reinserted during the selectedcontent projection process, bypassing the library's intended sanitization guarantees. No public exploit code has been identified at the time of analysis, and a vendor patch (3.11.0) is available; EPSS data was not provided but the CVSS 4.0 score of 5.1 reflects the user-interaction requirement limiting opportunistic exploitation.
Mutation XSS in justhtml (versions ≤1.11.0) allows injection of arbitrary HTML into serialized output when applications use a custom sanitization policy that retains raw-text elements such as <style> or <script>. Attacker-controlled text node content containing a closing tag sequence causes the serializer to break out of the raw-text context, inserting attacker-supplied markup into the final HTML string rendered by a victim's browser. No public exploit has been identified at time of analysis, but a vendor-released patch exists in version 1.12.0.
Mutation XSS (mXSS) in justhtml 1.13.0 and earlier enables markup injection bypass against applications using non-default custom SanitizationPolicy configurations that preserve foreign namespaces. Specially crafted input sanitizes to apparently clean markup that is subsequently reinterpreted as executable script when re-parsed by a browser - a classic parser-differential attack. No public exploit identified at time of analysis and this CVE is not in CISA KEV; however, the no-special-authentication-required network attack vector and the availability of a confirmed patch commit make prompt upgrading the correct response for any application using affected custom policies.
Cross-site scripting in justhtml versions before 1.13.0 allows attackers who control `<pre>` element content to break out of fixed-length Markdown code fences and inject raw HTML into the serialized output. When an application passes attacker-supplied HTML through `JustHTML(..., sanitize=True).to_markdown()` and a downstream CommonMark or GFM renderer processes the result with raw-HTML enabled, the injected script executes in the victim's browser. A proof-of-concept with full reproduction steps is publicly available in the GHSA advisory; no public exploit identified at time of analysis beyond that PoC.
Sanitization bypass in justhtml (Python HTML sanitization library) before version 1.17.0 enables cross-site scripting in applications using custom policies that preserve SVG or MathML namespaces, custom policies that preserve <style> blocks, or programmatic DOM construction prior to serialization. Crafted HTML input containing mutation-XSS parser-differential payloads, HTML integration points such as SVG <foreignObject> or MathML <annotation-xml encoding="text/html">, or resource-loading CSS constructs can survive sanitization and activate as executable markup after browser reparse. No public exploit code has been identified and the CISA KEV list does not list this CVE; notably, the default JustHTML(..., sanitize=True) configuration is not affected by most of the reported issues.
Missing capability check in WooCommerce Bookings WordPress plugin before 3.9.0 allows any authenticated user with Subscriber-level access or above to create draft bookable products via a vulnerable AJAX endpoint. The nonce check that would normally prevent unauthorized requests can be defeated by simply omitting the nonce token from the request - a particularly low-friction bypass. No public exploit has been identified at time of analysis, and the EPSS score of 0.19% (9th percentile) reflects low observed exploitation interest.
Broken access control in Brave Popup Builder WordPress plugin (versions through 0.8.5) allows any authenticated low-privilege user - including Subscriber-level or WooCommerce Customer accounts - to read restricted popup content by supplying an arbitrary post ID in the URL. The flaw is classified as CWE-639 (Authorization Bypass Through User-Controlled Key), an IDOR pattern where the server fails to validate whether the requesting user has rights to the resource identified by a user-supplied key. No active exploitation has been identified and EPSS sits at the 8th percentile, though the low barrier to exploitation (any authenticated user) means the attack surface on multi-tenant or membership-based WordPress sites is broader than the CVSS score alone suggests.
Session cookie forgery in Apache::AppSamurai::Util through version 1.01 is possible because CreateSessionAuthKey derives its entire entropy from observable runtime state - wall-clock microseconds and the Apache worker process ID - rather than a cryptographically secure random source. An unauthenticated remote attacker who can bound the second in which a target session was created and enumerate the small PID space can construct candidate 64-character hex cookies and recover a valid session token through online brute-force, bypassing authentication for all resources protected by the module. No public exploit code has been identified and SSVC confirms no current exploitation, though the attack is mechanically feasible given the constrained key space.
Privilege escalation in the Content Mask WordPress plugin (before 1.8.5.5) allows any authenticated Contributor to publish posts and pages without holding the publish capability, bypassing WordPress's role-based access control. The root cause is a missing capability check (CWE-269) when the plugin registers and handles its custom post type publishing workflow. No public exploit identified at time of analysis; EPSS is 0.18% (8th percentile), and the flaw is not listed in CISA KEV, reflecting negligible observed exploitation interest.