CVE-20XX-?
agent?

The internet is broken right now.

67 critical CVEs without a fix in the last 48 hours. Somewhere a staging server is public. It always is. 151 more already have a fix that someone still has to apply.

1,090 CVEs · updated 2026-11-19 10:22 UTC

Reset all
  1. CVE-2026-104639

    The Razorpay for WooCommerce WordPress plugin before 4.8.9 does not sanitize and escape a value derived from an incoming webhook request before using it in a SQL query, allowing users able to obtain the webhook secret (administrators) to perform SQL injection attacks.

    Injection Web app

    4.1 MEDIUM 0.2% exploit odds No fix listed
  2. CVE-2026-119288

    Zuke is a code-first, strongly typed build automation system for Deno. Prior to 1.51.0, @zuke/core launches resolved Windows .cmd and .bat shims by prepending cmd /c, causing cmd.exe to reparse a command line assembled with C-runtime quoting rules. A build running on Windows that resolves a tool from node_modules/.bin and passes attacker-influenced text containing an unquoted &, |, or ^ metacharacter can execute an additional attacker-chosen command in the build's context with access to its credentials. POSIX platforms are not affected. This issue is fixed in version 1.51.0.

    Injection Library

    7.0 HIGH 0.2% exploit odds Fix listed
  3. CVE-2026-117572

    The WP Visitor Statistics (Real Time Traffic) WordPress plugin before 8.8 does not sanitize and escape a parameter before using it in a SQL statement, allowing high-privilege users such as administrators to perform SQL injection attacks.

    Injection Web app

    4.1 MEDIUM 0.2% exploit odds No fix listed
  4. CVE-2026-86401

    Dell PowerProtect Data Domain Management Center, Feature Release versions 7.7.1.0 through 8.8, LTS2024 release versions 7.13.1.0 through 7.13.1.80, LTS2025 release versions 8.3.1.0 through 8.3.1.40, LTS2026 release versions 8.6.1.0 through 8.6.1.20, contain an Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability. A high privileged attacker with remote access could potentially exploit this vulnerability, leading to Information exposure.

    Injection

    2.7 LOW 0.2% exploit odds No fix listed
  5. CVE-2026-90719

    Dell PowerProtect Data Domain Management Center, Feature Release versions 7.7.1.0 through 8.8, LTS2024 release versions 7.13.1.0 through 7.13.1.80, LTS2025 release versions 8.3.1.0 through 8.3.1.40, LTS2026 release versions 8.6.1.0 through 8.6.1.20, contain an Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability. A high privileged attacker with remote access could potentially exploit this vulnerability, leading to Information disclosure.

    Injection

    2.7 LOW 0.2% exploit odds No fix listed
  6. CVE-2026-90723

    Dell PowerProtect Data Domain Management Center, Feature Release versions 7.7.1.0 through 8.8, LTS2024 release versions 7.13.1.0 through 7.13.1.80, LTS2025 release versions 8.3.1.0 through 8.3.1.40, LTS2026 release versions 8.6.1.0 through 8.6.1.20, contain an Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability. A high privileged attacker with remote access could potentially exploit this vulnerability, leading to Information disclosure.

    Injection

    2.7 LOW 0.2% exploit odds No fix listed
  7. CVE-2026-90724

    Dell PowerProtect Data Domain Management Center, Feature Release versions 7.7.1.0 through 8.8, LTS2024 release versions 7.13.1.0 through 7.13.1.80, LTS2025 release versions 8.3.1.0 through 8.3.1.40, LTS2026 release versions 8.6.1.0 through 8.6.1.20, contain an Improper Neutralization of Special Elements used in an SQL Command ('SQL Injection') vulnerability. A high privileged attacker with remote access could potentially exploit this vulnerability, leading to Information disclosure.

    Injection

    2.7 LOW 0.2% exploit odds No fix listed
  8. CVE-2026-129235

    Kubeshark is an eBPF-powered network observability tool for Kubernetes. From 53.1.0 until 53.4.0, the download_file tool in cmd/mcpRunner.go accepts caller-controlled path and dest values, fetches a selected Kubeshark Hub API response, and writes the returned bytes through os.Create(dest) in callDownloadFile without confining the destination. An attacker who influences arguments through an AI agent driving kubeshark mcp can write Hub-supplied bytes to any path writable by the local MCP user and can read captured network traffic from default auth-disabled Hub deployments. The tool is registered without --allow-destructive; reliable code execution is not demonstrated because the bytes are constrained to Hub API responses. Hub-side default authentication remains outside this remediation, so captured traffic can remain readable in default auth-disabled Hub deployments after the local destination-write fix. The destination-write issue is fixed in version 53.4.0.

    Injection Server

    5.3 MEDIUM 0.2% exploit odds Fix listed
  9. CVE-2026-13162

    An OS command vulnerability exists in CoreTec product that allows authenticated users with certain roles to execute arbitrary system commands with full control over the device. The vulnerability exists in the configuration daemon's time synchronization handler, which is responsible for setting the system time-zone. When processing time-zone configuration updates, the daemon constructs a shell command by concatenating user-supplied input without any sanitization or validation.

    Injection

    7.7 HIGH 0.2% exploit odds No fix listed
  10. CVE-2026-63496

    augment-vir is a collection of helper functions. Prior to version 31.73.5, the grep() function exported by @augment-vir/node inserts caller-controlled search patterns directly into a double-quoted shell command string without sanitization (grep.ts:310). When the constructed command is passed to spawn() with shell: 'bash' (shell.ts:98-99), bash evaluates $() command substitution expressions embedded in the pattern, resulting in arbitrary OS command execution as the Node.js process user. Any downstream application that passes untrusted input to grep()'s pattern or patterns field is vulnerable to unauthenticated local (and potentially remote) code execution. This issue has been patched in version 31.73.5.

    Injection Device

    7.8 HIGH 0.2% exploit odds Fix listed
  11. CVE-2026-115262

    The Duplicator WordPress plugin before 5.1.0, duplicator-pro WordPress plugin before 5.1.0 does not restrict its database charset and collation rewrite to schema statements during a restore, so a value a Subscriber stores in their own profile can break out of its string literal and inject SQL into the restore's INSERT statement when an administrator restores the backup onto a server lacking one of the source's charsets or collations.

    Injection Web app

    4.8 MEDIUM 0.2% exploit odds No fix listed
  12. CVE-2026-125672

    A flaw was found in cockpit-storaged. A Tang keyserver address read from a clevis token in a LUKS header is concatenated into curl's argument vector without a "--" separator. An address beginning with "-" is parsed by curl as an option instead of a hostname, allowing an attacker who can attach a crafted LUKS/clevis block device (for example, a local unprivileged user with an active console or desktop login session) to have an administratively-privileged Cockpit user unknowingly trigger curl option injection when editing the device's Tang keyserver settings. This can lead to arbitrary file read or write as the Cockpit session user.

    Injection

    6.6 MEDIUM 0.2% exploit odds No fix listed
  13. CVE-2026-121858

    The Event Tickets with Ticket Scanner WordPress plugin before 3.2.9 does not properly escape a search value before using it in SQL queries in several of its admin listings, allowing users with access to its admin area (administrators by default, or any role granted access in its settings) to perform SQL injection attacks.

    Injection Web app

    6.3 MEDIUM 0.2% exploit odds No fix listed
  14. CVE-2026-107264

    Improper Neutralization of Special Elements Used in a Template Engine vulnerability in Drupal Inline Formatter Field allows Code Injection. This issue affects Inline Formatter Field versions: from 0.0.0 to 4.2.0.

    Injection Web app

    5.4 MEDIUM 0.2% exploit odds No fix listed
  15. CVE-2026-107259

    Improper Neutralization of Special Elements Used in a Template Engine vulnerability in Drupal Linked Field allows Code Injection. This issue affects Linked Field versions: from 0.0.0 to 1.8.0.

    Injection Web app

    5.4 MEDIUM 0.2% exploit odds No fix listed
  16. CVE-2026-113644

    In Eclipse ioFog iofogctl versions 1.0.0 through 3.7.0 and 3.8.0-rc.1 through 3.9.1, the provisioning key and CA certificate returned by the Controller API are concatenated into shell commands (`edgelet provision` and `edgelet config cert` on supported releases; `iofog-agent provision` and `iofog-agent cert` on older releases) that run over SSH on a remote agent host and via a local shell. A party who can change that API response can include shell metacharacters and cause the operator's CLI to run additional commands on the agent host when the operator deploys or attaches an agent. This is fixed in version 3.9.2.

    Injection

    5.5 MEDIUM 0.2% exploit odds Fix listed
  17. CVE-2026-100101

    On Linux and other Unix-like systems, LibreOffice can send the current document as an email attachment with File > Send > Email Document, which starts the mail client through a script bundled with LibreOffice. The script let the shell interpret the name of the document's file, so a file with a crafted name could run commands when it was sent by email. In fixed versions the script passes the file name and the other details to the mail client as plain arguments.

    Injection

    8.4 HIGH 0.2% exploit odds No fix listed
  18. CVE-2026-65151

    NVIDIA NeMo Agent Toolkit contains vulnerability where an attacker could cause a SQL injection issue. A successful exploit of this vulnerability might lead to data tampering, denial of service and information disclosure.

    Injection

    8.4 HIGH 0.2% exploit odds No fix listed
  19. CVE-2026-123809

    Dynatrace MCP Server is an MCP server for Dynatrace Observability. Prior to 2.0.0, the create_workflow_for_notification tool inserts the caller-controlled teamName, problemType, and channel values into a Dynatrace workflow without neutralizing Jinja2 expressions. A caller can place template expressions such as event() in those values, and the workflow engine evaluates them when the persistent workflow runs. When the workflow has a valid Slack connector, the resulting Slack action can disclose event or tenant metadata to an attacker-selected channel; the shipped placeholder connection does not deliver messages by itself. The workflow remains in the tenant after the MCP session and credentials are gone. The isPrivate option defaults to false, making the created workflow tenant-visible unless the caller changes it. Exploitation requires either an MCP prompt-injection path or an operator approving a prompt that displays the template fragment literally. This issue is fixed in version 2.0.0.

    Injection Server

    4.2 MEDIUM 0.2% exploit odds Fix listed
  20. CVE-2026-103400

    An Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection') vulnerability in the ember service in Juniper Networks Session Smart Router and Session Smart Conductor allows an attacker with local access to unmount arbitrary filesystems or cause a denial of service via the creation of specially named directories and a service restart.

    Injection Network gear

    5.5 MEDIUM 0.2% exploit odds No fix listed
  21. CVE-2026-124686

    In Progress® Kendo® UI for jQuery versions prior to 2026.4.1111, kendo.init() compiles data-* attribute values with new Function when an invalid JSON object is supplied. An attacker with low privileges who can control these attribute values can inject JavaScript that executes during declarative widget initialization when a user loads or interacts with an affected page. Successful exploitation can compromise the confidentiality and integrity of data accessible in the user's browser context. The issue was reported in version 2022.3.1109; the first affected release has not been established.

    Injection Web app

    5.4 MEDIUM 0.2% exploit odds No fix listed
  22. CVE-2026-123535

    Tornado before 6.6 does not validate the max_age argument of RequestHandler.set_cookie (also called by set_signed_cookie, set_secure_cookie, and clear_cookie), allowing injection of additional Set-Cookie attributes. When an application passes attacker-influenced input as max_age, an attacker can embed attributes such as Domain, Path, or Secure to widen the cookie's scope to parent paths or other subdomains.

    Injection Web app

    2.3 LOW 0.2% exploit odds Fix listed
  23. CVE-2026-109830

    A vulnerability has been identified in meta-iot2050 (All versions >= V01.03.01). Affected devices take parts from a configuration file and execute them unchecked. This could allow an attacker which is able to manipulate the configuration file to make the user execute arbitrary code on the device.

    Injection Device

    7.0 HIGH 0.2% exploit odds No fix listed
  24. CVE-2026-65139

    NVIDIA Slurm contains an attacker-controlled SQL injection attack in certain SlurmDBD RPCs that can lead to arbitrary code execution.

    Injection Server

    6.3 MEDIUM 0.2% exploit odds No fix listed
  25. CVE-2026-66294

    XML injection in the Saxy XML encoder. Saxy.Encoder.cdata/1 serializes a {:cdata, characters} node by wrapping the supplied characters between <![CDATA[ and ]]> and passing the payload through byte for byte. The encoder's character-data and attribute-value paths escape the XML metacharacters <, >, &, ", and ', but the CDATA path applies no neutralization at all. XML provides no character escaping mechanism inside a CDATA section, so the only correct way to represent a ]]> substring is to split the section and rejoin it as ]]]]><![CDATA[>. Saxy neither performs that split nor rejects the substring. An application that renders untrusted text as CDATA, by building the node with Saxy.XML.cdata/1 and serializing it with Saxy.encode!/2 or Saxy.encode_to_iodata!/2, therefore lets an attacker terminate the section at a position of their choosing. Everything supplied after the ]]> is emitted outside any CDATA context and is parsed as markup, so a single attacker-controlled text field becomes attacker-controlled document structure: arbitrary sibling elements and attributes that the application never built. Downstream consumers of the serialized document accept the injected markup as genuine. Services that generate SOAP envelopes, RSS or Atom feeds, sitemaps, SVG, or XML exports from user-supplied content can have trusted fields forged or overridden. Applications that never place untrusted text in a CDATA node are not affected. This issue affects saxy: from 0.7.0 before 1.7.0.

    Injection Qcam Saxy

    2.1 LOW 0.2% exploit odds Fix listed
  26. CVE-2026-66295

    Saxy's XML encoder writes the target and the content of a processing instruction into the output document verbatim, without validation or escaping. The encoder builds the construct by concatenating both fields between <? and ?>, so neither field is checked against the constraints the XML specification places on it. Because ?> terminates a processing instruction, a target or content value containing that sequence closes the construct at a position the attacker chooses. Everything the attacker writes after it is emitted as ordinary markup at document level, which allows arbitrary elements, attributes, and further processing instructions to be injected into the serialized document. Appending a matching <? keeps the encoder's own terminator well-formed, so the result still parses and the injection is silent rather than a visible corruption. The target is additionally unconstrained by the XML Name production, which permits injection before the remainder of the processing instruction. The result is XML injection into a serialized document: the tree the application asks Saxy to encode and the tree a downstream parser reads back are no longer the same. Character data, attribute values, and comments are all escaped by the encoder, so processing instructions are the one construct that passes user bytes through untouched. Since Saxy.XML.processing_instruction/2 is a documented public builder, an application inherits the flaw without misusing the API. The consequences depend on what consumes the XML: forged records in a feed or export read by another system, injected stylesheet or rendering directives, or active content where the output is served to a browser as XML or SVG. This issue affects saxy: from 0.7.0 before 1.7.0.

    Injection Browser & mail Qcam Saxy

    2.1 LOW 0.2% exploit odds Fix listed
  27. CVE-2026-66293

    XML injection in the encoder of Saxy. Saxy.Encoder does not neutralize the -- sequence in XML comment content: escape_comment/2 returns the comment binary unchanged in every clause and only appends a space when the comment ends in a single -, so neither an interior -- nor a full --> is escaped or rejected before the encoder wraps the content in the <!-- and --> delimiters. When an application serializes attacker-controlled text as a comment, through Saxy.XML.comment/1 or a raw {:comment, text} simple form child, the attacker can close the comment early with --> and inject arbitrary sibling markup into the serialized document. A downstream parser then sees elements, attributes, and text that the application never built. Where the document is rendered as HTML or SVG, the injected markup executes as stored cross-site scripting; where it is consumed by another XML processor, the attacker can forge data that the producing application never emitted. Documents that Saxy only parses are not a vector. This issue affects saxy: from 0.7.0 before 1.7.0.

    Injection Library Qcam Saxy

    2.1 LOW 0.2% exploit odds Fix listed
  28. CVE-2026-115805

    Improper handling of local filesystem paths in Rapid7 Insight Agent on Linux and macOS may allow authenticated local users to gain elevated privileges. Exploitation requires existing local access and depends on the Agent's inventory collection processing the crafted path.

    Injection OS & kernel

    7.8 HIGH 0.2% exploit odds No fix listed
  29. CVE-2026-130928

    Code injection in Mobile in Google Chrome on on Android prior to 158.0.8127.36 allowed a local attacker leveraging social engineering to bypass web origin policy via a co-installed app. (Chromium security severity: Medium)

    Injection Browser & mail Google Chrome

    7.1 HIGH 0.2% exploit odds No fix listed
  30. CVE-2026-130378

    AdminNeo before 5.9.1 contains a filter injection vulnerability in its MongoDB driver that allows attackers to affect other documents by embedding ' AND ' or ') OR (' delimiters in a string _id or searched value. Attackers who can store a crafted document can cause a victim's delete or edit to match unintended documents, deleting the entire collection or modifying another document.

    Injection OS & kernel adminneo-org AdminNeo

    2.1 LOW 0.2% exploit odds Fix listed
  31. CVE-2026-81472

    Dell iDRAC Service Module (iSM) for Windows, Dell iDRAC Service Module (iSM) for Linux, and Dell iDRAC Service Module (iSM) for ESXi, versions prior to 6.1.1.0, contains an Improper Control of Generation of Code ('Code Injection') vulnerability. A high privileged attacker with local access could potentially exploit this vulnerability, leading to Elevation of privileges.

    Injection OS & kernel

    4.2 MEDIUM 0.2% exploit odds No fix listed
  32. CVE-2025-45528

    An issue in Nitro Software, Inc Nitro PDF Pro for macOS v.14.7 allows a local attacker to execute arbitrary code via a crafted PDF

    Injection

    7.3 HIGH 0.2% exploit odds No fix listed
  33. CVE-2026-118631

    Philips Hue Bridge Pro Argument Injection Restricted File Write Vulnerability. This vulnerability allows local attackers to modify files on affected installations of Philips Hue Bridge Pro. An attacker must first obtain the ability to execute low-privileged code within a container on the target system, in order to exploit this vulnerability. The specific flaw exists within the factory data service, which listens on TCP port 5001 by default. The issue results from the lack of proper validation of a user-supplied string before using it to execute a system call. An attacker can leverage this vulnerability to modify files in the context of the service account. Was ZDI-CAN-33090.

    Injection Device

    3.8 LOW 0.2% exploit odds No fix listed
  34. CVE-2026-107091

    An issue in HUION Inspiroy Frego S (L310) Windows Driver 20.0.1.118 allows a local attacker to execute arbitrary code via the HUION driver installer DLL-loading logic

    Injection OS & kernel

    7.8 HIGH 0.2% exploit odds No fix listed
  35. CVE-2026-123498

    Vikunja veans before 2.7.1 contains a prompt injection vulnerability that allows project members with write access to inject instructions by setting project titles or task content rendered verbatim inside the coding agent's authoritative instruction block. Attackers can craft ProjectTitle or task descriptions that the prime hook loads on SessionStart and PreCompact, causing the agent to modify the developer's checkout or run commands.

    Injection

    2.1 LOW 0.2% exploit odds Fix listed
  36. CVE-2026-111492

    SQL injection in the push notification handler of the MatchPoint Android app, which processes messages from Firebase Cloud Messaging and constructs queries by concatenating untrusted data. An attacker capable of interacting with the component could manipulate queries and access sensitive information from the local database, including credentials stored in the versions described in the report. The scope of the injection is limited to the device.

    Injection Device

    8.6 HIGH 0.2% exploit odds No fix listed
  37. CVE-2026-118378

    Tabularis before 0.28.0 contains an SQL injection vulnerability in its SQLite driver that allows MCP clients to execute stacked statements by supplying a crafted table_name to the describe_table tool. Attackers, including AI agents steered by untrusted content, can close the quoted PRAGMA argument and append statements to read, modify, or drop SQLite data, bypassing read-only mode and write approval.

    Injection OS & kernel

    8.5 HIGH 0.2% exploit odds Fix listed
  38. CVE-2026-108387

    radare2 is a UNIX-like reverse engineering framework and command-line toolset. Prior to 6.2.2, attacker-controlled data from several binary format plugins could reach RBinInfo->bclass. Opening or analyzing a crafted Pebble or Sega Mega Drive binary could place a dbg:// URI in bclass, causing radare2 to reopen the input through the debug input/output plugin and ultimately pass attacker-controlled arguments to execvp(). This could execute commands with the privileges of the radare2 process without an explicit debugging command. This issue is fixed in version 6.2.2.

    Injection Device

    8.5 HIGH 0.1% exploit odds Fix listed
  39. CVE-2026-129128

    Goose CLI 1.35.0 before 1.53.0 contains a local code execution vulnerability that allows attackers to run commands by setting diff.external in a repository's .git/config. Attackers can ship a crafted repository that, when the victim runs goose review, executes git diff without --no-ext-diff, running commands as the user and exposing credentials and provider API keys.

    Injection

    8.4 HIGH 0.1% exploit odds Fix listed
  40. CVE-2026-56838

    PraisonAI is a multi-agent teams system. From version 4.5.87 to before version 4.6.59, PraisonAI recipe execution blocks default-denied dangerous tools unless the caller explicitly passes allow_dangerous_tools=True. The normal recipe.run() path enforces this with _check_tool_policy(). The streaming path, recipe.run_stream(), loads the same recipe, checks dependencies, and then calls _execute_recipe() without running the dangerous-tool policy check. As a result, a recipe that honestly declares execute_command in TEMPLATE.yaml requires.tools is denied by recipe.run(), but reaches the execution engine through recipe.run_stream() with allow_dangerous_tools=False. This issue has been patched in version 4.6.59.

    Injection

    7.8 HIGH 0.1% exploit odds Fix listed
  41. CVE-2026-121547

    Canop dysk before 3.8.2 contains a terminal escape sequence injection vulnerability that allows local users to inject control sequences by crafting mount metadata rendered unsanitized in table and CSV output. Attackers can create FUSE mounts with ESC/OSC sequences in mount-point names or fsname, causing spoofed or hidden output and window-title changes when another user runs dysk.

    Injection

    1.0 LOW 0.1% exploit odds Fix listed
  42. CVE-2026-116576

    systeminformation is a System and OS information library for node.js. Prior to 5.33.16, services() in lib/processes.js inserts a caller-supplied service name into a PowerShell Get-CimInstance filter after processing it with sanitizeShellString() in lib/util.js. On Windows, the sanitizer removes ASCII quotes but permits Unicode quotation marks that PowerShell accepts as string delimiters, allowing crafted input to terminate the filter and place a PowerShell expression in the command executed by util.powerShell(). The expression runs with the Node.js process account's permissions and may read or change accessible files, alter application data, or disrupt the service. The package does not expose a network service itself, so remote reachability depends on a consuming application forwarding attacker-controlled input to services() or getDynamicData(). This issue is fixed in version 5.33.16.

    Injection Library

    7.8 HIGH 0.1% exploit odds Fix listed
  43. CVE-2026-130338

    AdminNeo before 5.9.1 contains a stored sql injection vulnerability in the MySQL driver that allows database users with table creation rights to inject SQL through unescaped geometry column names in unconvert_field(). Attackers can craft a column name containing an SQL expression that executes with another AdminNeo user's privileges when they edit a row, save it, or export the table.

    Injection OS & kernel adminneo-org AdminNeo

    2.1 LOW 0.1% exploit odds Fix listed
  44. CVE-2026-119282

    Nx is a monorepo solution for TypeScript and polyglot codebases. From 22.0.2 until 22.7.9 and 23.1.2 for @nx/maven, and from 21.0.0 until 22.7.9 and 23.1.2 for @nx/gradle, the plugins' project-graph analyzers launch through a shell and interpolate plugin options from nx.json into command arguments. The @nx/maven targetNamePrefix value and arbitrary @nx/gradle option names or values can therefore be interpreted as shell syntax whenever an nx command constructs the project graph. A user who runs Nx in an attacker-controlled repository or pull request can execute injected commands with the user's privileges, while workspaces that register neither affected plugin are not exposed to this path. Remediation requires upgrading both nx and each affected plugin to matching releases because the plugin fix depends on code in the nx package. This issue is fixed in versions 22.7.9 and 23.1.2.

    Injection Web app

    7.3 HIGH 0.1% exploit odds Fix listed
  45. CVE-2026-118905

    Node Version Manager (nvm) is a POSIX-compliant shell function for managing multiple node.js versions. In versions 0.31.2 through 0.40.5, `nvm_ls` filters installed versions with a sed program that embeds the requested version pattern in a `\#...#` address, escaping `.` (and, since 0.40.4, `#`) but not backslash. A version string containing `#` (or `\#`) therefore closes the address early, and the rest of it is parsed as sed commands: with GNU sed, the `e` command runs a shell command as the user; with BSD sed, the `w` and `r` commands write and read files. This is reachable whenever an attacker-influenced version string is passed to `nvm use`, `nvm which`, `nvm ls`, `nvm version`, `nvm run`, or `nvm exec`, or to `nvm-exec` via `NODE_VERSION`, and at least one version is installed. nvm's own `.nvmrc` parser strips comments, so a bare `nvm use` is not affected; however, the bash `cdnvm` and zsh `load-nvmrc` auto-switch hooks documented in the README read `.nvmrc` as-is and pass it to `nvm ls` and `nvm version`, so a user with one of those hooks installed runs the attacker's command just by changing into a directory under a malicious `.nvmrc`; scripts that run `nvm use "$(cat .nvmrc)"` are affected the same way. Version 0.40.6 passes any version pattern containing `#` or a newline through the `.nvmrc` content parser first, so no `#` reaches the sed program.

    Injection Library

    7.1 HIGH 0.1% exploit odds Fix listed
  46. CVE-2026-51831

    An issue in IDrive IDrive Backup for Windows Version 7.0.1.1 and below affected, remediated in Version 7.0.1.2 allows a local attacker to execute arbitrary code via the id_service.exe component

    Injection OS & kernel

    7.8 HIGH 0.1% exploit odds Fix listed
  47. CVE-2026-51832

    An issue in IDrive IDrive Backup for Windows Version 7.0.1.1 and below affected, remediated in Version 7.0.1.2 allows a local attacker to execute arbitrary code via the id_service.exe component

    Injection OS & kernel

    7.8 HIGH 0.1% exploit odds Fix listed
  48. CVE-2026-116958

    A flaw was found in konflux-ci. This vulnerability allows arbitrary code execution due to improper quoting of shell variables in deployment scripts. A local attacker who can supply arguments to the affected scripts could exploit this issue to trigger unintended command substitution, executing commands with the privileges of the deployment runner.

    Injection Server

    6.7 MEDIUM 0.1% exploit odds No fix listed
  49. CVE-2026-114354

    Missing authorization in the native binder service 'android.dji.server' (transaction code 1018) in DJI CrystalSky ZS600C firmware v0.3.9.7 through v0.6.6.9 on Android allows a local application with no permissions to execute arbitrary OS commands as root, because the handler passes a caller-supplied string to system() in the root-privileged 'assistant' daemon without any caller-identity or permission check.

    Injection Device

    8.6 HIGH 0.1% exploit odds No fix listed
  50. CVE-2026-87932

    LibreOffice Writer can save documents in DOCX format. A document can contain content controls whose values are bound to custom XML data, and that data is updated with the current values when the document is exported, including in a conversion run from the command line. Values taken from the document were not escaped where they were used, so a crafted document could write a file of its choosing anywhere the user can write when it was exported to DOCX. In fixed versions these values are written into the document directly, and the XSLT transformations LibreOffice runs can no longer write files or use the network.

    Injection

    6.8 MEDIUM 0.1% exploit odds No fix listed