A flaw in Node.js DNS resolution allows hostnames containing a NUL byte to be silently truncated by the dns.resolve*() functions and dns.reverse(). This is an incomplete fix of CVE-2026-48930, which only covered dns.lookup() and the net/tls host options.
Because the name is passed to c-ares as a C string, resolving attacker.example\0.trusted.example queries attacker.example and returns its records without any error. Applications that validate a hostname, for example against a domain allowlist, and then resolve it with dns.resolve*() can act on records for a different name than the one they validated.
This vulnerability affects Node.js 26.x, 24.x, and 22.x.
A flaw in Node.js HTTP Agent connection reuse can cause a client to accept as valid a response that was sent before the client has sent the request. This is an incomplete fix of CVE-2026-48931.
The guard that destroys idle keep-alive sockets when unexpected data arrives only runs when the event loop polls for I/O. If an unsolicited response has already reached the socket but has not been read yet when the socket is handed to a new request, for example because the event loop was busy with synchronous work, the guard is removed without seeing it and the data is parsed as the response to the new request.
This vulnerability affects Node.js **26.x**, **24.x**, and **22.x**.
Hugging Face Datasets through 5.1.0 contains an insecure deserialization vulnerability that allows attackers to execute arbitrary Python code by crafting a malicious pickle file loaded through the deprecated pandas packaged builder. Attackers can supply a pickle-backed dataset that is processed by Pandas._generate_tables via pandas.read_pickle() without validation, so when a victim calls datasets.load_dataset("pandas", data_files=...), the embedded __reduce__ payload executes code with the victim process's privileges.
An OS command injection vulnerability in the /htdocs/web/diagnostic.php endpoint of D-Link DAP1160L A1_FW101WWb04 allows attackers to execute arbitrary commands via injecting arbitrary shell metacharacters.
e-SIC Livre (esiclivre) v0.2.2 and earlier contains an uncontrolled resource consumption vulnerability in restrito/inc/captcha.php. The application uses user-supplied parameters (l, a, tf, ql) without validation to generate CAPTCHA images, allowing remote attackers to cause denial of service via excessive memory and CPU consumption.
Uni-CLI gives people and agents one command interface for websites, browsers, desktop applications, local tools, and MCP. Prior to version 0.225.2, Uni-CLI exposed the legacy JSON-RPC-over-HTTP MCP transport on loopback without validating browser Origin headers before routing requests. A malicious web page could send a CORS simple POST request, such as text/plain, to the local /mcp endpoint and deliver a JSON-RPC body to the dispatcher. If the user had started the local MCP HTTP transport, that page could drive tools/call requests against the user's local Uni-CLI server. The Streamable HTTP transport already enforced this browser-to-localhost boundary. The legacy stateless HTTP path did not, so the two HTTP transports had drifted. This issue is about the browser-to-localhost boundary; it does not change Uni-CLI's local-code-execution trust model. This issue has been patched in version 0.225.2.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 25.08 to before version 26.05, the IIIF Auth 2.0 endpoints auth/login and auth/token in AuthorizationFlowResource accept an origin query parameter without validation and then (1) embed it into the Content-Security-Policy: frame-ancestors response header and (2) embed it into an HTML response body as a JavaScript single-quoted postMessage argument. Because the attacker controls both the login and token calls that initiate the flow, the unescaped value yields reflected XSS. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 22.04 to before version 26.05, the unauthenticated REST endpoint GET /api/v1/index/{solrField}/spatial/search concatenates the user-supplied region query parameter directly into a Solr query string with no escaping, no local-param stripping, and no runtime schema enforcement (the OpenAPI @Schema(pattern=...) annotation is documentation-only at the JAX-RS layer). An attacker can inject Solr local-param syntax ({!parent ...}, {!join ...}) and boolean operators to make access-restricted georeferenced records match the query, bypassing the access-condition filter. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 4.11.0 to before version 26.05, several Jakarta WebSocket endpoints in goobi-viewer-core (DownloadTaskEndpoint, CampaignEndpoint, AutocompleteEndpoint, UserEndpoint) accept connections without verifying that the associated HttpSession belongs to a logged-in user, and none validate the Origin header against an allowlist. An unauthenticated attacker can therefore connect to these endpoints; through CampaignEndpoint this allows polluting the in-process crowdsourcing page-lock map and broadcasting forged lock messages, disrupting reviewers' work. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 25.01 to before version 26.05, AuthenticationEndpoint.verifyOpenIdToken builds its verifier as JWT.require(algorithm).withIssuer(issuer).build() and does not enforce the aud claim during cryptographic validation. The aud check is performed manually afterwards via a single string-equality call, which only handles single-string aud values. In a multi-RP IdP setup, a token whose single-string aud matches the viewer's client ID but was issued for a different relying party passes the check and authenticates the attacker as the token's sub. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. Prior to version 26.05, Configuration.getWebApiToken returns the empty string when webapi.authorization.token is not configured. AuthorizationFilter compares the supplied token against this default without first checking that the configured token is non-empty. A request to any @AuthorizationBinding-protected endpoint with an empty ?token= value is therefore accepted. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. Prior to version 26.05, the legacy RSS servlet (RssResolver, GET /rss/) accepts the q and filterQuery parameters and concatenates them into Solr queries without validation. The servlet predates the REST API layer and lacks the input validation present in GET /api/v1/records/rss. q forms the main query body (wrapped only in literal parentheses); filterQuery is passed directly as a Solr fq. An unauthenticated attacker can inject Solr boolean operators to break out of the wrapping and render the trailing access-condition filter-query suffix optional, so that access-restricted records are included in the generated RSS feed. The q and filterQuery parameters are the same underlying vulnerability in the same servlet; q (the main query body) has the broader scope. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 23.03 to before version 26.05, SolrTools.cleanUpQuery() strips Solr local-parameter syntax ({!...}) from user-supplied query strings before they reach Solr. Its regex can be bypassed by wrapping the local parameter in a second pair of curly braces ({{!...}}): the outer braces are stripped, the inner {!...} survives intact, and a valid Solr local parameter is forwarded. The flaw is in the shared sanitizer, so it affects any endpoint that routes user input through cleanUpQuery. The double-brace bypass applies from v23.10 (the greedy \{(.+)\} regex); a related bare-!join rewrap path is present. This issue has been patched in version 26.05.
Isar is an integration system for automated root filesystem generation. In versions 0.11 and newer, /etc/apt/trusted.gpg.d is created manually during bootstrap setup to be able to deploy local keys. By that, the directory is owned by the calling user (the one that executes Isar) instead of root. If the calling user's id is identical to one of an unprivileged user inside the image, this user is able to alter existing keys and deploy new ones, silently breaking the apt repo integrity protection. The issue is has no impact on those images that use an immutable rootfs, or if apt is never invoked. Changing only the keys does not trigger an exploit, it's just one step in a chain. An attacker would also have to manipulate a remote repo or intercept non-secure communication with it. This issue has been patched via commit 29e84ed.
Kitodo.Production is a workflow management tool for mass digitization and is part of the Kitodo Digital Library Suite. Prior to version 3.9.2, org.kitodo.production.helper.XMLUtils exposes several XML parsing helpers that construct a JAXP parser with the default, non-secure configuration. The methods parseXMLString(String), getNumberOfEADElements(String, String) and checkIfXmlIsWellFormed(String) build a DocumentBuilderFactory / XMLInputFactory / SAXParserFactory via newInstance() and parse the attacker-supplied string without disabling DOCTYPE declarations or external entity resolution. These helpers sit directly on the EAD-import path: an authenticated user who uploads an EAD/XML file through the "Create process" file-upload dialog has the uploaded bytes parsed by these methods. A crafted document with an external entity therefore yields XML External Entity (XXE) injection: local file disclosure and outbound SSRF requests from the server. This issue has been patched in version 3.9.2.
Conda Organization Infrastructure is a centralized location for infrastructure codes. Prior to commit 9c0f238, a GitHub Actions Pwn Request vulnerability appears to remain present in conda/conda. The affected workflow is .github/workflows/update.yml. It is triggered by issue_comment events containing @conda-bot render or @conda-bot recreate. For the render path, the workflow obtains the pull request head repository and branch from the PR API, checks out that PR-controlled repository/ref using secrets.SYNC_TOKEN, and then runs conda/actions/template-files@7f6830b1428a9bd47f0b068892c77eae95207037. The external template-files action reads its default configuration and templates from the checked-out working directory. In the confirmed finding, that action uses an unsandboxed Jinja2 environment to render attacker-controlled template content. This creates a Pwn Request chain in which an untrusted PR/comment path can reach template rendering in a privileged token context. This issue has been patched via commit 9c0f238.
Cacti is an open source performance and fault management framework. Prior to version 1.2.31, a User Enumeration vulnerability allows an attacker to identify valid usernames on the Cacti service. When a user attempts to log in, the application returns slightly different error messages depending on whether the provided username exists in the system. When username exists on Cacti, in the response, they see Access Denied! Login failed. When it does not exist on Cacti, they see Access Denied! Login Failed. The only difference between the responses is the capitalization of the letter "f" in the word failed. This issue has been patched in version 1.2.31.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 24.03 to before version 26.05, ExternalFilesDownloader.getFilename() extracts the destination filename from a server-supplied Content-Disposition response header and passes it directly to Path.resolve() without stripping directory components. This allows for a malicious HTTP server that the viewer fetches from to supply an absolute path or ../ traversal sequences as the filename, writing attacker-controlled content to an arbitrary path writable by the Tomcat process. When targeting the deployed web-application directory, this yields remote code execution as the tomcat OS user. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. From version 24.03 to before version 26.05, the DownloadTaskEndpoint WebSocket endpoint (/tasks/download/monitor.socket) in goobi-viewer-core accepts download requests from unauthenticated users. This issue has been patched in version 26.05.
Cacti is an open source performance and fault management framework. From version 1.2.27 to before version 1.2.31, Cacti contains an authenticated stored second-order SQL injection in the automation tree rule item handler. A user holding the Automation realm (default General Administration role) can persist an arbitrary SQL fragment in the automation_tree_rule_items.field column. The fragment is later concatenated directly into a SELECT ... AS source FROM ... clause in lib/api_automation.php :: create_all_header_nodes() and executed when an administrator triggers automation via the standard graphs.php?action=actions&drp_action=8 ("Apply Automation Rules") bulk action. The result of the injected SELECT is written into graph_tree_items.title and rendered in the Cacti tree UI, providing a direct exfiltration channe. This issue has been patched in version 1.2.31.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. Prior to version 26.05, when a password reset link is visited, the application generates a replacement password in a predictable way and emails it to the account holder. The insufficient entropy allows the password to be brute-forced by an attacker who can approximate the server-side generation time. Exploiting this finding in isolation requires the victim to have triggered the reset themselves and the attacker to infer the server time from network or email metadata. This issue has been patched in version 26.05.
The Goobi viewer is a web application that allows digitised material to be displayed in a web browser. Prior to version 26.05, the password reset token is predictably based on the time of the reset request. An attacker who knows the target's email address can request a reset and then brute-force the token. Exploiting this finding alone forces an unauthorized password reset on any account. This issue has been patched in version 26.05.
Cacti is an open source performance and fault management framework. Prior to version 1.2.31, the prior blocklist ([;&| `$\\n\r]) omitted single-quote, double-quote, redirect operators, subshell parens and braces. An authenticated admin could save an input_string` containing single-quote, redirect operator, or subshell parens, and pass the check. This issue has been patched in version 1.2.31.
obsidian-tc is a governed, agent-ready Model Context Protocol server for Obsidian vaults. From 1.0.1 until 1.31.4, `create_entity`, `link_entities`, `rename_entity`, `unlink_entities` and `delete_entity` regenerate memory notes from SQLite at the computed `memory/<type>/<name>.md` path without reliably preserving note ownership or the consistency of a failed lifecycle operation. During shared materialization, a hand-authored note or an orphan left by a different entity whose name sanitized to the same path could have its entire body overwritten on the next write naming that type and name, with non-owned frontmatter preserved. `rename_entity` also seeded the destination path with a direct atomic write without validating ownership and could overwrite a foreign note there. Separately, several tools mutated SQLite before fallible filesystem work: `create_entity` and `link_entities` inserted the row or relation before materialization; `rename_entity` updated the entity, including a retired status, before updating its note; `unlink_entities` removed a relation before rematerializing the source note; and `delete_entity`'s cascade removed the target and its relations before rematerializing neighbouring notes. A later filesystem failure could therefore leave a row or relation changed and the memory graph inconsistent with the notes on disk even when the operation failed. An authorized caller with the required memory-tool permissions could, without intending to, destroy the body of a vault note outside the entity being written; `delete_entity` additionally requires its deletion permission and destructive confirmation, which this memory flaw does not bypass. Snapshots-before-write through `restore_note` mitigate the overwrite only where snapshots are enabled and the note had previously been written through obsidian-tc. This issue is fixed in version 1.31.4.
obsidian-tc is a governed, agent-ready Model Context Protocol server for Obsidian vaults. From 1.13.0 until 1.31.4, on a Streamable HTTP deployment with `auth.jwtSecret` configured and a 2026-07-28-era client declaring `elicitation.form` support, a dispatch-gated tool call offered an `inputRequired` confirmation could still complete after the human declined or cancelled, or the elicitation response was omitted or malformed, provided the client echoed the validly bound `requestState` on resubmission. The server trusted an authentic, call-bound state as proof of approval even though that state was minted before the `elicitation/create` leg was asked and proved only that the resubmission belonged to the same call. A client behaving incorrectly or under an attacker's control could therefore execute dispatch-gated destructive vault operations without the human's approval, while still requiring the caller's ordinary tool permissions. Only tools marked `destructive: true` or requiring a scope with a human-in-the-loop floor were reachable through this bypass. The sixteen handler-side conditionally gated tools, including `write_note` overwrite and `move_note`, did not read elicitation state on the modern wire and let neither approval nor decline through that path. Affected deployments using stdio or leaving `auth.jwtSecret` unset instead used the CLI `elicit_token` fallback and did not offer this `requestState` round trip. The fix does not prevent a malicious client from reporting a fabricated approval, because the server cannot verify that the client displayed the form or that a human saw it. This issue is fixed in version 1.31.4.
GFI Archiver before 15.14 contains a stored cross-site scripting vulnerability in the MailInsights scheduled report configuration that allows authenticated attackers to inject arbitrary web script or HTML via the report name parameter to /Archiver/MailInsights.aspx. The injected payload is stored by ReportScheduling.btnSaveReport_Click() without output encoding and is executed in the browser of the user who created the scheduled report when they subsequently view the MailInsights page.
An issue in @justice8096/sagemath-mcp-server 1.0.1 allows a remote attacker to execute arbitrary code via the escapeShellString and sanitizeCode functions.
A local privilege escalation vulnerability in the optional dnscache package of Zimbra Collaboration Suite (ZCS) allows the Unix zimbra account to access files with root privileges. The passwordless sudo policy permits arbitrary arguments to zmunbound, whose wrapper forwards them to Unbound without restricting the configuration path. An attacker can supply a configuration that retains root privileges and disables the chroot. Include directives expose root-readable file contents through parser errors, while an attacker-selected logfile path creates a root-owned file or appends Unbound log records to an existing file. This can disclose password hashes and private keys and corrupt privileged configuration files, preventing affected services from starting with the modified configuration. The affected configuration is present in ZCS 10.1.16 and 10.1.21 dnscache packages.
A missing authorization vulnerability in Zimbra Collaboration Suite (ZCS) allows a delegated administrator with no explicit administrative grants to overwrite global Zimlet configuration through ConfigureZimletRequest. The uploaded configuration can also add an attacker-controlled domain to the proxy allowlist of the default class of service. If a signed-in user in that class opens a crafted /service/proxy URL, Zimbra serves attacker-controlled HTML under its own origin. The resulting JavaScript can read mailbox data and perform authenticated actions as the victim. Exploitation requires a delegated-administrator account and access to the admin SOAP service; the browser compromise additionally requires the victim to open the crafted URL. ZCS 10.1.16 and 10.1.20 are affected.
A reflected cross-site scripting vulnerability in the login page of Zimbra Collaboration Suite (ZCS) allows an unauthenticated attacker to execute JavaScript through the zrememberme parameter. A crafted relogin URL selects a two-factor error display that inserts the parameter into a hidden input attribute without HTML encoding. A victim who opens the URL receives attacker-controlled script and page content under the legitimate Zimbra origin, allowing credential capture through a replacement login form. The relogin path logs out an existing session before rendering the payload, so the demonstrated attack does not inherit that session's mailbox access. ZCS 10.1.16 and 10.1.17 are affected.
An LDAP filter injection vulnerability in Zimbra Collaboration Suite (ZCS) allows an authenticated user to query otherwise unreturned attributes of calendar resources through SearchCalendarResourcesRequest. Zimbra escapes search values but inserts the supplied attribute names directly into LDAP filter expressions without validation. An attacker can inject additional conditions and use repeated matching queries to recover readable resource attributes such as internal mailbox hostnames and mail transport routes. The search remains restricted to calendar-resource objects and the LDAP service account's access rights. ZCS 10.1.16 and 10.1.21 are affected.
A cross-account authorization vulnerability in Zimbra Collaboration Suite (ZCS) allows an authenticated user to insert forged shared-file records into another user's mailbox through FileSharedWithMeRequest. The request bypasses delegated-account authorization and accepts attacker-selected file metadata, owner identifiers and declared sizes without validating a corresponding document share. An attacker can plant entries attributed to another real local account without that account sharing a document. No victim interaction is required. ZCS 10.1.16 and 10.1.20 are affected.
A CSS sanitization vulnerability in Zimbra Collaboration Suite (ZCS) allows an authenticated user with Briefcase access to publish a document that displays attacker-controlled content under the Zimbra origin. When processing a stylesheet containing non-ASCII characters, DefangFilter appends the final ASCII segment without sanitizing it. HTML stored as application/x-zimbra-doc can therefore retain CSS URL references that the normal sanitizer removes. An attacker can share a crafted document that loads external resources and displays a credential-collection form using Zimbra's own visual assets. Exploitation requires the victim to open the document in a browser that renders the content type inline; credential theft additionally requires form submission. ZCS 10.1.16 and 10.1.20 are affected.
A cross-site request forgery vulnerability in the OAuth extension of Zimbra Collaboration Suite (ZCS) Network Edition allows an attacker to link an attacker-controlled external account to a victim's mailbox. The OAuth callback does not bind its state value or authorization code to the initiating Zimbra session and instead selects the destination mailbox from the browser's current authentication cookie. On a deployment with the provider configured, an attacker can induce a signed-in victim to load a callback containing a fresh code for the attacker's provider account. This creates a persistent data source that imports attacker-controlled contacts or events. With bidirectional CalDAV synchronization, appointments subsequently created or edited in the linked calendar are also disclosed to the attacker. ZCS 10.1.16 and 10.1.20 are affected.
A hostname canonicalization vulnerability in the email HTML sanitizers of Zimbra Collaboration Suite (ZCS) allows an unauthenticated sender to bypass the restriction on forms that target the Zimbra host. DefangFilter and the OWASP ActionAttributePolicy compare the form-action hostname with the request hostname before browser URL canonicalization. A percent-encoded hostname can therefore survive sanitization while the browser submits the form to the normal Zimbra origin with the recipient's authentication cookie. This permits authenticated form requests from a crafted email after the recipient opens it in Classic and submits the form. The issue is present in ZCS 10.1.16 and 10.1.21. It does not bypass an endpoint's CSRF-token checks; a state-changing attack additionally requires an endpoint that accepts the resulting request without that protection.
An organizer authentication vulnerability in Zimbra Collaboration Suite (ZCS) allows an unauthenticated email sender to insert calendar invitations attributed to another user. The email delivery path permits external invitations but does not verify that the message sender is authorized to use the iCalendar ORGANIZER identity. When automatic invitation insertion is enabled, a crafted REQUEST component creates an event containing attacker-selected meeting details and reminders under a forged organizer without recipient interaction. No corresponding event is created in the claimed organizer's calendar. ZCS 10.1.16 and 10.1.21 are affected.
An attendee authentication vulnerability in the email calendar processing of Zimbra Collaboration Suite (ZCS) allows an unauthenticated sender to forge another attendee's response to an existing meeting. Zimbra passes a null sender identity when processing iCalendar REPLY components, causing the attendee authorization filter to be skipped. An attacker who knows the meeting UID, organizer address and attendee address can send a non-outdated reply that changes the attendee's participation status in the organizer's calendar. The change occurs automatically during delivery without organizer or attendee interaction. ZCS 10.1.16 and 10.1.21 are affected.
An organizer authentication vulnerability in the email calendar processing of Zimbra Collaboration Suite (ZCS) allows an unauthenticated sender to cancel or modify an existing meeting in a recipient's calendar. Zimbra accepts iCalendar CANCEL and REQUEST components without verifying that the email sender is authorized to act as the supplied organizer. An attacker who knows the meeting UID and organizer address can supply a non-outdated sequence number to move the recipient's meeting to Trash or replace its details. Processing occurs automatically during email delivery without recipient interaction. ZCS 10.1.16 and 10.1.21 are affected.
Entire CLI hooks into Git workflows to capture AI agent sessions. Prior to version 0.7.7, a path traversal vulnerability in Entire CLI allows an attacker with push access to the checkpoints repository to craft malicious checkpoint metadata that causes entire session resume or entire checkpoint rewind to write attacker-controlled transcript data outside of the expected session directory. The issue occurs because checkpoint metadata is fetched from the remote entire/checkpoints/v1 branch and the SessionID field was used to construct filesystem paths without validation in the restore path. A malicious SessionID containing absolute paths or path traversal sequences could cause arbitrary files on the victim’s machine to be overwritten. This issue has been patched in version 0.7.7.
NetBox MCP Server is a simple read-only Model Context Protocol server for NetBox. Prior to version 1.2.1, when running with TRANSPORT=http, the MCP server has no authentication on the HTTP endpoint. Any client that can reach the port gets a valid MCP session and full access to all four NetBox tools - no token or credential needed. This issue has been patched in version 1.2.1.
The GTM Server Side plugin by stape.io implements server-side tracking on WordPress websites. From version 2.0.0 to before version 2.1.48, the WooCommerce purchase event tracking mechanism allows an unauthenticated attacker to trigger the rendering of another customer's full order data — including personally identifiable information (PII) — into the page's JavaScript dataLayer. This issue has been patched in version 2.1.48.
odhcpd is the DHCPv6, DHCPv4, and router-advertisement server used by OpenWrt. In versions 25.12 and prior, odhcpd's DHCPv6 self-loop detection at dhcpv6.c:720 uses a network-byte-order DUID length value directly as the memcmp() size parameter. On little-endian hosts, this causes memcmp() to read far beyond the stack buffer — a single unauthenticated DHCPv6 packet triggers a 2560-byte overread from a 128-byte buffer. At time of publication, there are no publicly available patches.
odhcpd is the DHCPv6, DHCPv4, and router-advertisement server used by OpenWrt. Prior to version 25.12, odhcpd's DHCPv6 IA response handler writes past the end of a fixed 512-byte stack buffer (pdbuf) when a single DHCPv6 REQUEST carries enough IA options to nearly exhaust the reply space. Two independent overflow sites exist in the same call chain, both reachable from the network via a single unauthenticated UDP packet. odhcpd runs as root on OpenWrt; embedded targets often lack stack canaries and ASLR, making code execution realistic. This issue has been patched in version 25.12.
odhcpd is the DHCPv6, DHCPv4, and router-advertisement server used by OpenWrt. From version 17.01.0 to before version 25.12.3, odhcpd's DHCPv6 IA parser accepts an IA_NA or IA_PD option whose body is shorter than struct dhcpv6_ia_hdr, casts the option header to that full structure, and reads ia->iaid before proving that the IA body contains the IAID field. If an unauthenticated network-adjacent client places a zero-length or otherwise short IA option at the end of odhcpd's 8192-byte receive stack buffer, the IAID read crosses from data_buf into adjacent stack memory. The disclosed bytes are then reflected in the DHCPv6 reply IAID; prior ARM validation showed data_buf[8192:8196] disclosing the saved stack canary, and wider stack scans can leak ASLR-useful stack/process contents. This issue has been patched in version 25.12.3.
odhcpd is the DHCPv6, DHCPv4, and router-advertisement server used by OpenWrt. Prior to version 25.12, odhcpd's DHCPv6 IA handler keeps a per-message `first` lease pointer so multiple IA options in one message reuse the same Reconfigure authentication key. A DHCPv6 REQUEST from an adjacent-network client can make the first IA set `first`, then make a later same-IAID IA fail the on-link check and free that same lease, leaving `first` dangling. A following IA in the same packet allocates a new lease and copies `first->key`, producing a 16-byte heap-use-after-free read. ASan confirms the UAF; native tracing shows the packet shape can deterministically reclaim the freed lease as the destination allocation, producing a zeroed Reconfigure key copy. The validated impact is unauthenticated adjacent-network denial of service and DHCPv6 Reconfigure/ForceRenew key confusion. Arbitrary PC control is not proven for this UAF alone. This issue has been patched in version 25.12.
A local privilege escalation vulnerability in the Zimbra Collaboration Suite (ZCS) MTA package allows the Unix zimbra account to execute arbitrary commands as root. The installed sudo policy permits passwordless root execution of Postfix administration commands without restricting their arguments. An attacker can select a zimbra-controlled daemon_directory in the Postfix configuration and invoke postfix set-permissions, causing Postfix to execute the attacker's post-install program as root. The Postfix configuration directory is also writable by zimbra, so restricting postconf alone does not remove the execution path. Exploitation requires local command execution as zimbra on an MTA node. ZCS 10.1.16 and 10.1.21 are affected.