mammoth (aka mammoth.js) before 1.12.2 is vulnerable to prototype pollution when reading the styles defined in a document. Converting a crafted .docx file allows an attacker to add arbitrary properties to Object.prototype. In 1.11.0 through 1.12.1, applications that convert further documents in the same process and return the converted HTML can also disclose the contents of local server files (to the party supplying the documents) by setting externalFileAccess to true.
RabbitMQ is a messaging and streaming broker. Prior to versions 3.13.15, 4.0.20, 4.1.11, and 4.2.6, The management UI uses EJS 1.0 in which <%= ... %> does NOT HTML-escape. connection.ejs:135 renders <%= connection.ssl_details.peer_cert_subject %> (and peer_cert_issuer) directly into the page. The same pattern appears in streamConnection.ejs:102,106,110. The values come from rabbit_ssl:peer_cert_subject/1 which formats the DN as a string without HTML escaping. The verifier corrected the original researcher's claim: this is reachable only when the listener is configured with verify_peer (so the certificate must be signed by a CA in the broker's trust store, not arbitrary self-signed); however, in deployments using mTLS for client authentication, any user who can request a certificate from the organisational CA controls the Subject CN. An attacker who can obtain a TLS client certificate signed by a CA the broker trusts (with verify_peer enabled) can embed JavaScript in the certificate's Subject DN. When any administrator views that connection in the management UI, the script executes in the admin's browser session, allowing full account takeover (create users, export definitions, etc.). The management UI's CSP includes 'unsafe-inline', so inline script execution is not blocked. Preconditions include TLS listener configured with ssl_options.verify = verify_peer Attacker can obtain a CA-signed client certificate with attacker-chosen Subject (e.g. self-service corporate PKI, or rabbitmq_trust_store plugin in use) Administrator views the connection detail page. This issue is fixed in versions 3.13.15, 4.0.20, 4.1.11, and 4.2.6.
IBM Financial Transaction Manager (FTM) for RedHat OpenShift is vulnerable to RAG poisoning via unauthenticated runbook upsert (CWE-74) in the FTM AI agent server (api.vectordb.runbooks.js:51). An unauthenticated attacker can insert malicious runbook content into the agent's vector database to steer AI-driven MCP tool calls, potentially triggering unauthorized payment actions or exfiltrating payment data.
MCP Atlassian is a Model Context Protocol (MCP) server for Atlassian products (Confluence and Jira). Prior to 0.22.0, the OAuth fallback token directory and JSON file are created without explicit owner-only modes. Local users or processes with access through the resulting group or world permission bits can read access and refresh tokens and reuse the associated Atlassian session. The advisory traces the vulnerable input and processing flow through ~/.mcp-atlassian, oauth-<client_id>.json, access_token, and refresh_token, which identify the affected entry points, controls, and code paths. This issue is fixed in version 0.22.0.
9Router is an AI router & token saver. Prior to 0.5.6, 9Router deployments that allow requests to reach Next.js without the sanitizing custom-server.js wrapper use the client-supplied X-9r-Real-Ip value as the bucket key in getClientIp, checkLock, and recordFail in src/lib/auth/loginLimiter.js for POST /api/auth/login. A remote unauthenticated attacker can rotate the header on every password guess so each request uses a new failed-attempt bucket and the five-attempt progressive lockout never returns HTTP 429. This permits unthrottled password guessing against the dashboard login and can lead to an administrative session if the password is recovered. This issue is fixed in version 0.5.6.
9Router is an AI router & token saver. Prior to 0.5.6, 9Router deployments that allow requests to reach Next.js without the sanitizing custom-server.js wrapper trust the client-supplied X-9r-Real-Ip header in src/dashboardGuard.js when isLocalRequest decides whether canAccessPublicLlmApi may skip API-key validation for /api/v1/* routes. A remote unauthenticated attacker can set X-9r-Real-Ip to 127.0.0.1 and be classified as a local client, including on the verified GET /api/v1/models route. This permits unauthorized use of the instance owner's configured LLM providers, consumption of paid credits, and enumeration of configured providers and models. This issue is fixed in version 0.5.6.
Novu provides an API for sending notifications through multiple channels. Prior to 3.18.0, Novu accepts chat webhook URLs from subscriber credentials.webhookUrl, channel endpoint endpoint.url, event payload.webhookUrl, and event overrides.webhookUrl, then passes the selected endpoint.url through send-message-chat.usecase.ts to raw HTTP requests in the Slack, Discord, Mattermost, Microsoft Teams, Grafana On-Call, Ryver, Rocket.Chat, GetStream, and Zulip providers. An authenticated user can supply an internal or otherwise restricted destination because these paths do not apply normalizeOutboundHttpUrl, assertSafeOutboundUrl, or the DNS-pinned safeOutboundJsonRequest protection used by the generic webhook providers. The Novu worker can consequently issue attacker-directed POST requests to internal network services and cause interactions or actions supported by those services. This issue is fixed in version 3.18.0.
CVE-2026-89407 is a high-severity vulnerability in FasterXML jackson-core affecting versions from 2.17.0 through 3.2.1 in specified subranges. It involves inefficient regular expression complexity in the NumberInput.looksLikeValidNumber() method, causing quadratic time complexity when processing certain numeric string inputs. This can lead to denial-of-service conditions by exhausting server resources during JSON deserialization of numeric types. The issue does not exist before version 2.17.0 and has been fixed by replacing the problematic regex with a single-pass scan.
Improper Validation of Specified Quantity in Input in ZenHive mpp allows a client holding an open payment channel to obtain paid resources without being charged.
MPP.Session.Actions.accept_voucher/3 in lib/mpp/session/actions.ex treats a voucher whose cumulativeAmount equals the channel's already-accepted cumulative amount as an idempotent success, returning the channel unchanged without calling maybe_spend/2. The credential verifies, the protected resource is served, and spent and units stay where they were. Because the server issues a fresh challenge per request and the credential replay store keys on challenge id and payload, the same signed voucher can be re-presented under every new challenge, so one paid voucher yields an unbounded number of paid units. The path is reachable from any method built on MPP.Session.Method through the Plug, MCP, JSON-RPC and WebSocket transports.
This issue affects mpp: from 0.14.0 before 0.16.2.
The Product Feed Manager for WooCommerce – CTX Feed – Support 220+ Shopping & Social Channels plugin for WordPress is vulnerable to Directory Traversal in all versions up to, and including, 6.6.43 via the 'provider' parameter. This makes it possible for authenticated attackers, with shop manager-level access and above, to delete arbitrary files on the server, which can lead to remote code execution when critical files are deleted. Exploitation requires two sequential REST API calls: first to /wp-json/ctxfeed/v1/make_feed/save_feed_config to persist the traversal payload in wp_options, then to /wp-json/ctxfeed/v1/manage_feeds/delete_feed to trigger the unlink(); deletion is further constrained to files whose extensions match the plugin's validated whitelist (csv, xml, tsv, xls, xlsx, json, txt).
Joplin is an open source note-taking and to-do application that organises notes and lists into notebooks. Prior to 3.6.15 and 3.7.2, packages/renderer/MdToHtml/rules/fountain.ts passes HTML generated by the vendored fountain.js renderer into note output without sanitizing it. A malicious Fountain code block can therefore execute script when Fountain rendering is enabled in desktop or mobile clients, or when a note is published through Joplin Server where Fountain rendering is enabled by default. The script can read content subsequently loaded in the reused note viewer or, when published notes are served from the same domain as server content, access data available to an authenticated browser in the server origin. This issue is fixed in versions 3.6.15 and 3.7.2.
Sync-in Server versions prior to 2.4.0 contain a vulnerability in the sync diff endpoint where user-supplied strings are compiled into regular expressions without complexity validation. This can lead to catastrophic backtracking patterns that block the Node.js event loop, causing the server to become unresponsive until restarted. The issue is patched in version 2.4.0.
draw.io is a configurable diagramming and whiteboarding application. Prior to version 30.3.8, src/main/java/com/mxgraph/online/Utils.java checks IPv6 Unique Local Addresses in Utils.sanitizeUrl() by comparing the text prefixes fc00:: and fd00::, but the JDK returns the expanded address form, so the fc00::/7 range, including the AWS metadata range fd00:ec2::/32, is not blocked. An unauthenticated request to /embed2.js?fetch= can therefore make src/main/java/com/mxgraph/online/EmbedServlet2.java fetch an IPv6 ULA internal resource and reflect the response to the requester. Utils.validatedAddress() uses the same private-address check for the separate ProxyServlet path, which requires ENABLE_DRAWIO_PROXY=1. The primary /embed2.js path requires no proxy feature flag or DNS rebinding, and it can disclose cloud metadata credentials or data from other IPv6-reachable internal services. This issue is fixed in version 30.3.8.
hono before 4.13.7 fails to HTML-escape plain strings rendered by hono/jsx as a child or fallback of Suspense, as a string child of ErrorBoundary alongside an asynchronous sibling, as the single child of a Context.Provider, or as the root value passed to renderToString() or renderToReadableStream() from hono/jsx/dom/server. These paths stringify their input and treat the result as already-escaped markup, so an attacker who controls such a string during server-side rendering can inject arbitrary HTML and execute script under the application's origin.
The npm package 'chai-as-indexed' version 7.2.8 contains malicious code that exfiltrates the full environment variables of the installer's Node.js process to a remote server. The package's main entry point loads a module that sends process.env data to a base64-encoded URL, then executes arbitrary code returned from that server in the importing process context. This behavior allows remote code execution on every load of the package. The package masquerades as a legitimate chai plugin, and the command-and-control server URL is obfuscated using base64 encoding to evade detection.
Svelte devalue is a JavaScript library that serializes values into strings when JSON.stringify isn't sufficient for the job. In versions 5.1.0 through 5.9.2, stringify and uneval functions serialize a typed array by emitting its entire backing ArrayBuffer rather than only the view, so serializing a Node Buffer, whose backing store is a process-wide shared pool, discloses up to 64 KB of unrelated process memory, including bytes from other in-flight requests. In a server-side-rendered framework such as SvelteKit or Nuxt, a public page whose load() returns a small Buffer, or that reads a small file, can therefore ship another user's request body or Authorization header in its HTML without authentication. Because this occurs during serialization, it fires on every such render and is not mitigated by the parse/unflatten prototype-pollution and denial-of-service guards, which only apply when parsing untrusted input. As a workaround, convert Node Buffer objects to Uint8Array before serialization. This issue has been fixed in version 5.9.3.
PJSIP pjproject versions 2.17 and earlier have an improper certificate validation vulnerability in the OpenSSL and GnuTLS backends. The flaw involves incorrect handling of DNS SubjectAltName values containing embedded NUL bytes, allowing an attacker with a specially crafted certificate from a trusted issuer to impersonate a target server during TLS verification. This can lead to interception of SIP sessions and theft of REGISTER credentials. The mbedTLS backend is not affected. No fixed version is currently available.
The npm package test890-auth version 1.0.0 contains malicious code that executes during installation. It collects sensitive environment information and the full contents of the consuming project's package.json file, then sends this data to a hardcoded external server. The package does not provide any legitimate functionality and is designed as a dependency-confusion probe. Systems with this package installed should be considered fully compromised.
WeaselBiscuit is a JavaScript infostealer malware distributed through 11 malicious npm packages. It executes a detached Node.js process, retrieves payloads from Npoint URLs, and communicates with a command-and-control server at 103.170.217.184:8787. The malware profiles infected hosts, steals Chrome extension storage related to wallet signing states, captures clipboard contents, and logs Windows keystrokes on command. It is a stripped-down variant of DPRK-linked BeaverTail and OtterCookie malware families, lacking wallet draining, password decryption, screenshot capture, and remote shell capabilities. Numeric campaign identifiers in package names are used for infection tracking. Attribution to DPRK is suggested but not definitive.