CVE-20XX-?
agent?

CVEs of 2026.

Every CVE published in 2026 that we hold, ranked by exploit odds.

16,668 CVEs · updated 2026-11-20 20:25 UTC

Reset all
  1. CVE-2026-126185

    A improper control of generation of code ('code injection') vulnerability in Fortinet FortiAnalyzer 8.0.0 through 8.0.1, FortiAnalyzer 7.6.0 through 7.6.7, FortiAnalyzer 7.4.0 through 7.4.11, FortiAnalyzer Cloud 8.0.0 through 8.0.1, FortiAnalyzer Cloud 7.6.0 through 7.6.7, FortiAnalyzer Cloud 7.4.1 through 7.4.11, FortiManager 8.0.0 through 8.0.1, FortiManager 7.6.0 through 7.6.7, FortiManager 7.4.0 through 7.4.11, FortiManager Cloud 8.0.0 through 8.0.1, FortiManager Cloud 7.6.0 through 7.6.7, FortiManager Cloud 7.4.1 through 7.4.11 may allow an authenticated administrator to execute unauthorized code or commands via specially crafted requests.

    Injection Network gear

    7.2 HIGH 0.6% exploit odds No fix listed
  2. CVE-2026-126307

    Nezha Dashboard before 2.6.0 contains multiple authorization, tenant isolation, and resource exhaustion vulnerabilities that allow members or unauthenticated attackers to abuse NAT, DDNS, OAuth, cron, and server transfer features. Attackers can claim undeclared Dashboard hostnames to capture login bodies, hijack DDNS profiles, create executable schedules without exec scope, or exhaust memory via buffered NAT requests.

    Auth & access Web app

    8.7 HIGH 0.6% exploit odds Fix listed
  3. CVE-2026-63505

    Probo is a self-hostable governance, risk, and compliance (GRC) platform built for engineering and security teams. Prior to version 0.223.1, Probo's console GraphQL API is a multi-tenant GRC platform where all data is scoped to an organization. A Finding can reference a Risk via finding.riskId. That reference is stored without validating that the risk belongs to the caller's organization, and the Finding.risk field resolver authorizes the finding (caller's org) instead of the risk, then loads the risk with a tenant scope derived from the risk's own global ID. As a result, an authenticated member of organization A can attach an arbitrary organization B risk GID to a finding they own and then read organization B's Risk (name, description, treatment, category, severity, owner) — with no membership in B. The stored cross-tenant reference also blocks B from deleting its own risk (ON DELETE RESTRICT), a cross-tenant integrity/DoS. This issue has been patched in version 0.223.1.

    Auth & access

    4.9 MEDIUM 0.6% exploit odds Fix listed
  4. CVE-2026-121289

    On affected platforms running Arista EOS with BGP configured, a specially crafted packet can cause the BGP agent to crash. Successful exploitation causes the BGP agent process to terminate, resulting in the loss of all BGP sessions on the affected device and possible disruption to network routing until the process recovers. Due to the nature of the vulnerability, it may also be possible for a sophisticated attacker to achieve arbitrary code execution, which could compromise the confidentiality and integrity of the affected system.

    Memory safety

    9.3 CRITICAL 0.6% exploit odds No fix listed
  5. CVE-2026-121664

    Improper neutralization of input during web page generation ('cross-site scripting') in Visual Studio Code allows an unauthorized attacker to execute code over a network.

    XSS Microsoft Visual Studio Code

    8.3 HIGH 0.6% exploit odds Fix listed
  6. CVE-2026-124210

    Server-side request forgery (ssrf) in Microsoft Office SharePoint allows an authorized attacker to disclose information over a network.

    Disclosure Server Microsoft SharePoint Server Subscription Edition

    6.5 MEDIUM 0.6% exploit odds Fix listed
  7. CVE-2026-126966

    SmarterTools SmarterMail Missing Authorization Privilege Escalation Vulnerability. This vulnerability allows remote attackers to escalate privileges on affected installations of SmarterTools SmarterMail. Authentication is required to exploit this vulnerability. The specific flaw exists within the svcDomainAdmin.asmx endpoint. The issue results from the lack of authorization prior to allowing access to functionality. An attacker can leverage this vulnerability to escalate privileges to resources normally protected from the user. Was ZDI-CAN-33566.

    Auth & access SmarterTools SmarterMail

    8.8 HIGH 0.6% exploit odds No fix listed
  8. CVE-2026-128046

    OpenMage magento-lts 20.17.0 before 20.18.2 and 21.0.0-beta1 before 21.0.0-beta3 contains an unrestricted file upload vulnerability because the general/file/protected_extensions blocklist omits phar, php8, and phps. Unauthenticated attackers can upload a .phar file through a product File custom option to a predictable media/custom_options path and request it directly to execute code on servers that map .phar to PHP and serve that directory.

    Code execution Web app

    9.2 CRITICAL 0.6% exploit odds Fix listed
  9. CVE-2026-109218

    The Contact Manager plugin for WordPress is vulnerable to Arbitrary File Upload in all versions up to, and including, 9.1.2 via the add_message function. This is due to an incomplete extension blacklist that fails to detect dangerous extensions when a trailing dot causes the extension-derivation routine to return an empty string, which sanitize_file_name() then silently removes before the file is written to disk. This makes it possible for unauthenticated attackers to upload files that may be executable, which makes remote code execution possible. Successful code execution additionally requires the server to use Apache with libapache2-mod-php (which maps .phar to PHP by default) and the attacker to retrieve the uploaded file within the brief window before it is removed by unlink() at end of request.

    Code execution Web app

    9.8 CRITICAL 0.6% exploit odds No fix listed
  10. CVE-2026-76938

    Authentication Bypass Vulnerability in JP1/Automatic Job Management System 3. This issue affects JP1/Automatic Job Management System 3 - Manager: from 10-00 before 13-50-31.

    Auth & access

    9.8 CRITICAL 0.6% exploit odds No fix listed
  11. CVE-2026-82117

    Authentication Bypass by Spoofing Vulnerability in HiRDB. This issue affects HiRDB Server Version 10: from 10-00 before 10-12-01; HiRDB Server Version 9: from 09-00 through 09-66-22; HiRDB Server with Additional Function Version 9: from 09-00 through 09-66/V; HiRDB Server Version 9(32): from 09-02 through 09-66-22; HiRDB Server with Additional Function Version 9(32): from 09-00 through 09-66-/V.

    Auth & access Server

    9.8 CRITICAL 0.6% exploit odds No fix listed
  12. CVE-2026-111422

    In apache-airflow-providers-ssh before 7.0.1, SSHRemoteJobOperator quoted the values of its templated `environment` parameter only once, although they are placed inside an already-quoted wrapper script, so shell syntax such as `$(...)` in a value was executed on the remote SSH host. In a Dag that passes trigger-time input such as `params` or `dag_run.conf` into `environment`, a user who can trigger that Dag but cannot author Dags could run arbitrary commands on the remote host as the SSH connection's user. Users are advised to upgrade apache-airflow-providers-ssh to 7.0.1 or later.

    Injection Server

    8.8 HIGH 0.6% exploit odds Fix listed
  13. CVE-2026-67543

    Heym builds and runs agentic systems and business-process automations on self-hosted infrastructure. Prior to version 0.0.70, heym hardens one untrusted-code path but not its sibling. The Python tool executor (python_tool_executor.py) treats user code as untrusted and mandates a Docker sandbox (no network, read-only fs, non-root, no docker socket), failing closed. The skill executor (skill_python_executor.py) runs skill Python via a local backend subprocess with no Docker isolation — only a 512 MB RLIMIT and a 7-key env denylist — so a skill executes inside the backend execution context, with filesystem access, network egress, and non-denylisted environment variables that should not be exposed to untrusted/shared skills. In the shipped Docker config (backend runs as root + /var/run/docker.sock mounted), this escalates to host root. This issue has been patched in version 0.0.70.

    Injection Server

    8.8 HIGH 0.6% exploit odds Fix listed
  14. CVE-2026-126183

    A authorization bypass through user-controlled key vulnerability in Fortinet FortiManager 8.0.0 through 8.0.1, FortiManager 7.6.0 through 7.6.7, FortiManager 7.4.0 through 7.4.11, FortiManager 7.2.0 through 7.2.12, FortiManager Cloud 8.0.1, FortiManager Cloud 7.6.0 through 7.6.7, FortiManager Cloud 7.4.1 through 7.4.11, FortiManager Cloud 7.2.1 through 7.2.12 may allow an authenticated attacker to execute commands outside of the expected scope via specially crafted HTTP request.

    Auth & access Network gear

    8.7 HIGH 0.6% exploit odds No fix listed
  15. CVE-2026-115483

    The Everest Forms plugin for WordPress is vulnerable to Arbitrary File Read in all versions up to, and including, 3.6.3 via the attach_entry_files() function. This is due to the function attaching any entry field value whose meta key matches /signature_/ to the notification email after only a file_exists() check, without validating or restricting the path. This makes it possible for unauthenticated attackers to read the contents of arbitrary files on the server, such as wp-config.php, which can contain sensitive information. Successful exploitation requires a form whose email notification has the “Attach Files” setting enabled and that contains a field with a meta key matching /signature_/ (for example, a text field labelled “Signature”), with the notification routed to an attacker-supplied email address so the exfiltrated file is delivered to the attacker. Note: The Everest Forms Pro add-on is required to exploit this vulnerability, as it provides the signature field and the email file-attachment functionality (file-email-attachments) that the attack relies on.

    Disclosure Web app

    7.5 HIGH 0.6% exploit odds No fix listed
  16. CVE-2026-54161

    Network UPS Tools is a collection of programs which provide a common interface for monitoring and administering UPS, PDU and SCD hardware. From 2.8.3 to 2.8.5, upsmon in clients/upsmon.c inserts the server-supplied ups.alarm value into shell commands executed by notify(), async_notify(), and wall(). When a monitored device reports an ALARM state and an operator has configured NOTIFYCMD with EXEC in NOTIFYFLAG ALARM, shell metacharacters in the alarm text are evaluated on POSIX or Windows and can execute arbitrary operating-system commands as the upsmon account, which is commonly root. An attacker can control the value through a malicious or compromised upsd, a man-in-the-middle on a plaintext upsd to upsmon connection, or a compromised UPS device whose driver supplies the alarm upstream of upsd; TLS between upsd and upsmon does not mitigate the compromised-device path. This vulnerability is fixed in 2.8.6.

    Injection

    8.1 HIGH 0.6% exploit odds Fix listed
  17. CVE-2026-85899

    Use after free in Microsoft Edge (Chromium-based) allows an authorized attacker to elevate privileges over a network.

    Memory safety Browser & mail Microsoft Edge (Chromium-based)

    8.1 HIGH 0.6% exploit odds Fix listed
  18. CVE-2026-112546

    An out-of-bounds read was addressed with improved bounds checking. This issue is fixed in iOS 26.7.2 and iPadOS 26.7.2, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2, tvOS 27.2, visionOS 27.2, watchOS 27.2. Processing a maliciously crafted file may lead to unexpected app termination.

    Memory safety Device Apple Ipados

    7.5 HIGH 0.6% exploit odds Fix listed
  19. CVE-2026-79852

    A vulnerability in the web-based management interface of AOS could allow an unauthenticated remote attacker to conduct SQL injection attacks against the AOS instance. Successful exploitation could allow an attacker to execute arbitrary commands on the underlying operating system.

    Injection Web app

    9.8 CRITICAL 0.6% exploit odds No fix listed
  20. CVE-2026-79854

    A vulnerability in the affected interface of AOS-10 could allow an unauthenticated remote attacker to conduct SQL injection attacks against the AOS-10 instance. Successful exploitation could allow an attacker to execute arbitrary commands on the underlying operating system.

    Injection

    9.8 CRITICAL 0.6% exploit odds No fix listed
  21. CVE-2026-120779

    Improper control of generation of code ('code injection') in Visual Studio Code allows an unauthorized attacker to execute code over a network.

    Injection Microsoft Visual Studio Code

    9.6 CRITICAL 0.6% exploit odds Fix listed
  22. CVE-2026-38600

    A Zip Slip vulnerability in the /upload endpoint of codeskyblue gohttpserver through v1.3.0 allows attackers to write arbitrary files via uploading a crafted Zip archive.

    Disclosure

    9.8 CRITICAL 0.6% exploit odds No fix listed
  23. CVE-2026-112449

    An integer overflow was addressed with improved input validation. This issue is fixed in iOS 26.7.2 and iPadOS 26.7.2, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2, visionOS 27.2. Processing a maliciously crafted 3D model may lead to memory corruption.

    Memory safety OS & kernel Apple Ipados

    7.5 HIGH 0.6% exploit odds Fix listed
  24. CVE-2026-106156

    In Progress® Telerik® UI for ASP.NET Core versions `2015.1.318` through `2026.3.811`, repeated server-side aggregation requests, with or without grouping, cause dynamic assemblies and load contexts to be retained, resulting in accumulating process memory consumption. An unauthenticated remote attacker can exploit this if the application exposes an aggregation endpoint that is reachable without authentication. Repeated requests can cause memory growth that may exhaust available resources and affect application availability. Disabling grouping in the client-side interface does not prevent direct requests containing these descriptors. The issue is fixed in version `2026.3.1015`.

    Denial of service Web app

    8.6 HIGH 0.6% exploit odds Fix listed
  25. CVE-2026-110468

    WidzialniCMS is vulnerable to Unrestricted File Upload in the /panel/controllers/save_photo.php endpoint. Due to a client-controlled permission check parameter and insufficient validation of the currentpath parameter, an unauthenticated attacker can upload arbitrary files to any locations on the server, leading to Remote Code Execution. This issue has been fixed in version 2.5.

    Code execution Web app

    10.0 CRITICAL 0.6% exploit odds Fix listed
  26. CVE-2026-123298

    In JetBrains Hub before 2026.2.52556 sMTP command injection via email verification was possible

    Injection

    5.0 MEDIUM 0.6% exploit odds No fix listed
  27. CVE-2026-91115

    Silverstripe Framework is a PHP framework which powers the Silverstripe CMS. From 6.0.0 until 6.2.13, the writer option of i18nTextCollectorTask is passed to the dependency injector without validating that it names an allowed writer implementation. A crafted dev task URL can therefore instantiate arbitrary objects with attacker-controlled constructor arguments and execute code on the server. An administrator can be induced to run the task by following a link because the nested task did not require confirmation, and a site in dev mode exposes the path without authentication. CMS 5 is not affected. This issue is fixed in version 6.2.13.

    Code execution Web app

    8.8 HIGH 0.6% exploit odds Fix listed
  28. CVE-2026-116336

    Operation on a dropped index in PostgreSQL REINDEX CONCURRENTLY allows an object creator to execute arbitrary code as the operating system user running the database, via an index expression that issues DROP INDEX. Versions before PostgreSQL 18.7, 17.12, 16.16, 15.20, and 14.25 are affected.

    Memory safety Server

    8.8 HIGH 0.6% exploit odds No fix listed
  29. CVE-2026-116326

    Uncaught exceptions in PostgreSQL plperl allow an object creator to cause use of a stale sigjmp_buf, via a tied Perl object, or a Perl object whose overloaded conversion throws while plperl is handling another error. We have not ruled out arbitrary code execution on some platform. Versions before PostgreSQL 18.7, 17.12, 16.16, 15.20, and 14.25 are affected.

    Code execution

    8.8 HIGH 0.6% exploit odds No fix listed
  30. CVE-2026-116467

    A security flaw has been discovered in ggml-org llama.cpp up to 0.7.0. Affected by this vulnerability is an unknown functionality of the file src/llama-sampler.cpp of the component Mirostat Sampler. The manipulation results in reachable assertion. It is possible to launch the attack remotely. The project was informed of the problem early through a pull request but has not reacted yet.

    Other Library

    6.9 MEDIUM 0.6% exploit odds Fix listed
  31. CVE-2026-122164

    Weak authentication in Windows Kerberos allows an authorized attacker to elevate privileges over a network.

    Auth & access OS & kernel Microsoft Windows 10 1607

    8.0 HIGH 0.6% exploit odds Fix listed
  32. CVE-2026-109476

    Improper encoding or escaping of output in ASP.NET Core allows an unauthorized attacker to disclose information over a network.

    Other Microsoft .NET 10.0 Installed On Linux

    6.1 MEDIUM 0.6% exploit odds Fix listed
  33. CVE-2026-121544

    Koodo Reader through 2.4.6 contains a path traversal vulnerability in the self-hosted HTTP server's OPDS handlers that allows authenticated attackers to read files outside the upload directory. Attackers can upload a crafted books.db with a cover value containing ../ sequences, then request /opds/cover/{key} to read files accessible to the server process.

    Disclosure Server

    6.9 MEDIUM 0.6% exploit odds Fix listed
  34. CVE-2026-129961

    LMS (Lan Management System) is vulnerable to Remote Code Execution in its e-mail-based Helpdesk ticket intake. The mail parser does not reject Apache .htaccess configuration files or PHP-executable files when saving ticket attachments. A remote unauthenticated attacker can e-mail a Helpdesk ticket with such an attachment; in the default configuration, anyone can request the uploaded file's URL which executes attacker-supplied code with the privileges of the web server process. No operator interaction or LMS client session is required. The issue was fixed in commit d5059b1.

    Code execution Web app LMS

    9.5 CRITICAL 0.6% exploit odds Fix listed
  35. CVE-2026-126971

    KiCad Project Archive File Handling Directory Traversal Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of KiCad. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file. The specific flaw exists within the handling of project archive files. The issue results from the lack of proper validation of a user-supplied path prior to using it in file operations. An attacker can leverage this vulnerability to execute arbitrary code in the context of the current user. Was ZDI-CAN-33212.

    Disclosure KiCad

    7.8 HIGH 0.6% exploit odds Fix listed
  36. CVE-2026-126972

    KiCad Plugin Directory Traversal Remote Code Execution Vulnerability. This vulnerability allows remote attackers to execute arbitrary code on affected installations of KiCad. User interaction is required to exploit this vulnerability in that the target must visit a malicious page or open a malicious file. The specific flaw exists within the handling of plugin packages. The issue results from the lack of proper validation of a user-supplied path prior to using it in file operations. An attacker can leverage this vulnerability to execute arbitrary code in the context of the current user. Was ZDI-CAN-33085.

    Disclosure Web app KiCad

    7.8 HIGH 0.6% exploit odds Fix listed
  37. CVE-2026-129860

    Microsoft jxrlib through 1.1 and the mircomir jxrlib fork before 1.3.7 contain a reachable assertion vulnerability in ParsePFDEntry() that allows attackers to abort applications via an invalid orientation tag. Attackers can supply a crafted JPEG XR file with an out-of-range WMP_tagTransformation value to crash applications using default assertion-enabled builds, or store an unchecked orientation value otherwise.

    Denial of service Microsoft Jxrlib

    6.8 MEDIUM 0.6% exploit odds Fix listed
  38. CVE-2026-128129

    AIOHTTP is an asynchronous HTTP client/server framework for asyncio and Python. Prior to 3.14.4, aiohttp buffers all additional connection data received while a request marked for upgrade, such as a WebSocket handshake, is being handled, without a size limit or flow control. A remote attacker can stream data behind an upgrade request to a handler that does not immediately accept or reject the upgrade and cause unbounded memory consumption. This issue is fixed in version 3.14.4.

    Denial of service Server

    8.2 HIGH 0.6% exploit odds Fix listed
  39. CVE-2026-128339

    ExcelJS through 4.4.0 contains a memory exhaustion vulnerability in Worksheet._parseMergeCells that allows attackers to crash applications by supplying unbounded mergeCell ranges. Attackers can craft a small workbook with a mergeCell ref like A1:XFD1048576, causing xlsx.load() to create billions of Row and Cell objects and abort the process.

    Denial of service Library Exceljs Project Exceljs

    8.7 HIGH 0.6% exploit odds Fix listed
  40. CVE-2026-126378

    Vvveb through 1.0.8.6 contains a file upload vulnerability in the theme installer validFile() function that uses strpos() instead of strict path matching. Attackers with admin credentials can upload a ZIP containing a PHP file whose path includes /theme.php to achieve remote code execution on nginx deployments.

    Code execution Web app

    9.4 CRITICAL 0.6% exploit odds Fix listed
  41. CVE-2026-124057

    A security flaw has been discovered in TOTOLINK CA300-PoE 6.2c.884. This impacts the function setWiFiMultipleConfig of the file /cgi-bin/cstecgi.cgi. Performing a manipulation of the argument Key1Str1 results in buffer overflow. The attack can be initiated remotely. The exploit has been released to the public and may be used for attacks.

    Memory safety Network gear

    8.5 HIGH 0.6% exploit odds No fix listed
  42. CVE-2026-112688

    A memory corruption issue was addressed with improved memory handling. This issue is fixed in iOS 26.7 and iPadOS 26.7, iOS 27.2 and iPadOS 27.2, macOS Golden Gate 27.2, macOS Sequoia 15.8.2, macOS Tahoe 26.7.2. Parsing a file may lead to an unexpected app termination or arbitrary code execution.

    Memory safety OS & kernel Apple Ipados

    8.8 HIGH 0.6% exploit odds Fix listed
  43. CVE-2026-97631

    EDK2 contains a vulnerability in BIOS where a user may cause “Out-of-bounds Write”, Classic Buffer Overflow, and “Incorrect Calculation of Buffer Size” by remote access. Successful exploitation of this vulnerability can lead to arbitrary code execution. It could also impact Confidentiality, Integrity, and Availability.

    Memory safety OS & kernel

    9.5 CRITICAL 0.6% exploit odds Fix listed
  44. CVE-2026-128344

    ExcelJS through 4.4.0 contains a memory exhaustion vulnerability that allows attackers to crash applications by supplying a workbook with an unbounded image anchor column in the drawing part. Attackers can set an anchor column such as 300000000 in xl/drawings/drawing1.xml so reading getImages() anchor positions makes Worksheet.getColumn allocate millions of Column objects and abort with a heap out-of-memory error.

    Denial of service Exceljs Project Exceljs

    8.7 HIGH 0.6% exploit odds Fix listed
  45. CVE-2026-76954

    An improper neutralization of special elements used in an SQL command vulnerability in the Auth API in Synology DiskStation Manager (DSM) before 7.2.1-69057-13, 7.2.2-72806-10, 7.3.2-86009-5, and 7.4.1-90080-1 allows remote attackers to read or write arbitrary files and conduct denial-of-service attacks.

    Injection

    9.8 CRITICAL 0.6% exploit odds No fix listed
  46. CVE-2026-116947

    Heap-based buffer overflow in SQL Server allows an authorized attacker to execute code over a network.

    Memory safety Server Microsoft SQL Server 2017 (GDR)

    7.5 HIGH 0.6% exploit odds Fix listed
  47. CVE-2026-93011

    A flaw was found in dnsmasq. After a reduction of internal domain-name buffer sizes, the DNSSEC validator can write a few bytes past a name buffer when converting names to and from DNS wire format. An attacker who can cause the server to validate a crafted DNSSEC response may overwrite a small amount of adjacent heap metadata. DNSSEC support is not compiled in by default and must also be enabled at runtime.

    Memory safety Server

    5.3 MEDIUM 0.6% exploit odds No fix listed
  48. CVE-2026-113773

    Improper access control in Windows DHCP Server allows an unauthorized attacker to perform spoofing over a network.

    Auth & access OS & kernel Microsoft Windows Server 2016

    5.3 MEDIUM 0.6% exploit odds Fix listed
  49. CVE-2026-123776

    RRD::CGI::Image versions through 0.01 for Perl pass CGI query parameters containting "tz" to tzset(), , so a FIFO or device path blocks the CGI process. The normalize_params method sends any query key matching /tz/i to tz(), which sets $ENV{TZ} to that value, with no validation. It then calls tzset(). glibc treats a TZ value starting with / or : as a file path and opens it, so a value naming a FIFO or a blocking device stalls the process inside tzset().

    Other

    9.8 CRITICAL 0.6% exploit odds No fix listed
  50. CVE-2026-129280

    Multiple ServerUI and ServerUI2 SOAP operations in N-able N-central before 2026.4.1 HF3 do not verify the caller's session, permissions or customer scope. A remote attacker, in some cases without authentication, can read device and patch data, test passwords of user accounts, send messages to managed agents, and modify device and service settings belonging to other customers.

    Auth & access

    9.3 CRITICAL 0.6% exploit odds No fix listed

CVE records published in 2026 (ids starting CVE-2026-), sorted by how likely they are to be exploited.