🚨 CVE-2026-76234
libcrux-ecdh and libcrux-ed25519 before 0.0.6, and libcrux-psq before 0.0.7, contain cryptographic implementation bugs. libcrux-ecdh did not properly check length and clamping during X25519 secret validation (and had a broken clamping check for imported X25519 secret keys); libcrux-ed25519 performed a duplicated clamping step during key generation; and libcrux-psq panicked instead of propagating an AEADError. These were fixed in the respective patched releases.

🎖@cveNotify

🚨 CVE-2026-76237
stigmem-node before 0.9.0a12 contains a broken object level authorization (cross-tenant BOLA) vulnerability in the quarantine review endpoints. On multi-tenant deployments running the opt-in stigmem-plugin-multi-tenant, the list/count queries and _get_quarantined_fact in routes/quarantine.py lacked a tenant_id predicate and the garden lookup was not tenant-scoped, allowing a tenant administrator with only a plain tenant write capability to list, read, and admit or reject quarantined facts belonging to other tenants via the /v1/quarantine endpoints. Default single-tenant deployments are not affected.

🎖@cveNotify

🚨 CVE-2026-76239
Stigmem before 0.9.0a11 fails to validate the delivery_address parameter when creating webhook subscriptions, allowing authenticated users to specify internal loopback and private network destinations. Attackers can trigger matching fact-change events to cause the Stigmem server to issue server-side HTTP POST requests to internal services, enabling blind SSRF attacks against localhost and private network endpoints.

🎖@cveNotify

🚨 CVE-2026-75583
keeper.sh's calendar module version prior to 2.18.14 contains a server-side request forgery (SSRF) guard bypass vulnerability that allows authenticated attackers to reach private network addresses by exploiting a DNS rebinding attack against the two-phase URL validation and connection flow. The SSRF guard validates a hostname's resolved IP addresses but discards them before the actual HTTP connection is opened, allowing an attacker who controls authoritative DNS to return a public address during validation and a private address during the subsequent independent socket-level DNS resolution, causing the guard to pass while the outbound connection reaches internal infrastructure such as cloud instance metadata endpoints.

🎖@cveNotify

🚨 CVE-2026-61518
ISPConfig contains an authenticated SQL injection vulnerability in the Remote API. The primary_id parameter passed to delete and update API methods is concatenated directly into SQL WHERE clauses without integer casting or parameterized query binding. The built-in SQL injection scanner does not block quote-free boolean payloads and does not reject requests in its default configuration. A remote API user holding any single low-privilege function permission can inject arbitrary SQL to delete or modify records across all tenants in the control panel database and extract arbitrary data via blind boolean inference, including password hashes and client records.

🎖@cveNotify

🚨 CVE-2026-63722
ICEcoder 8.1 contains an unauthenticated remote code execution vulnerability that allows unauthenticated attackers to execute arbitrary OS commands by chaining an authentication bypass, CSRF validation bypass, and unsanitized command execution. Attackers can send a single HTTP POST request to the terminal endpoint with a password parameter to bypass authentication, a non-empty csrf parameter to skip CSRF validation, and an arbitrary command string passed directly to proc_open() to achieve remote code execution as the web-server user.

🎖@cveNotify

🚨 CVE-2026-76832
Agno's PythonTools in libs/agno/agno/tools/python.py contains a path traversal vulnerability that allows attackers to read, write, or execute arbitrary files by supplying parent-directory traversal sequences in the file_name argument passed to read_file, save_to_file, or run_python_file tool actions. Attackers can inject traversal sequences such as '../../../../../../etc/passwd' through direct tool invocation or via prompt injection embedded in agent-processed content to escape the intended base_dir boundary and achieve arbitrary file read, arbitrary file write, or arbitrary Python code execution within the process user's authority.

🎖@cveNotify

🚨 CVE-2026-77066
The scanFeedsResolver in packages/api/src/resolvers/subscriptions/index.ts passes the caller-supplied url straight to axios.get(url, rssParserConfig()) with no address validation. The same file guards the subscribe path with validateUrl(), which rejects private and reserved ranges through the private-ip library, and createPageSaveRequest applies the same check, so the omission is specific to this resolver. An authenticated user can direct the server to request arbitrary internal endpoints. The response is parsed as a feed or as HTML and the resolver returns the resulting url, title, description and type fields, so disclosure is limited to feed-shaped metadata and to link elements advertising RSS or Atom feeds; requests that do not parse still distinguish reachable ports from unreachable ones through the resulting error.

🎖@cveNotify

🚨 CVE-2026-77067
The setWebhookResolver in packages/api/src/resolvers/webhooks/index.ts stores the caller-supplied url without any address validation, and the file imports no validation helper. When a subscribed event fires, callWebhook in packages/api/src/jobs/call_webhook.ts issues axios.request with that url, the method and Content-Type recorded on the webhook, and a JSON body carrying the event data, so an authenticated user can make the server send repeated attacker-shaped requests to internal endpoints, including link-local metadata addresses. The request is blind: callWebhook discards the result and writes only a success line or the axios error to the server log, so the response is not returned through the API.

🎖@cveNotify

🚨 CVE-2026-76833
@cgauge/yaml npm package contains an arbitrary code execution vulnerability that allows attackers to execute arbitrary JavaScript by embedding a custom !js YAML tag whose construct callback unconditionally calls eval() on attacker-supplied string values during document parsing. Any application parsing untrusted YAML input with this library exposes full Node.js runtime authority, including environment variable access, filesystem read/write, network access, and subprocess execution, with no safe-mode alternative or opt-out mechanism available.

🎖@cveNotify

🚨 CVE-2026-75140
jsoup through 1.23.2, fixed in commit 862ba2f, contains an uncontrolled resource consumption vulnerability in XmlTreeBuilder that allows remote attackers to exhaust JVM heap memory by supplying a deeply nested XML document with uniquely-namespaced elements. The builder copies the entire inherited namespace map on every start element, causing quadratic time and memory complexity, which attackers can exploit to trigger an OutOfMemoryError and terminate the application.

🎖@cveNotify

🚨 CVE-2026-72847
broot renders each file and directory name in its interactive tree view exactly as read from the filesystem. Names are converted with a plain to_string_lossy() call in src/tree_build/builder.rs and in TreeLine::unprune in src/tree/tree_line.rs, and no control-character filtering exists anywhere in the code, even though the doc comment on the TreeLine name field states that some characters may have been stripped. Any local user who can create a file can therefore place an escape sequence in its name and have it written unmodified to the terminal of anyone who browses that directory, between broot's own styling codes. A reported proof of concept used an OSC 52 clipboard-write sequence and captured the raw bytes broot wrote to its pty, confirming the sequence reaches the terminal unstripped. What an injected OSC or CSI sequence can then do depends on the terminal emulator in use. Browsing a directory is broot's primary function and carries no expectation that the content is trusted.

🎖@cveNotify

🚨 CVE-2026-76641
Expat through 2.8.3 contains an out-of-bounds read vulnerability that allows attackers to trigger memory corruption by processing XML with external entity parsers created via XML_ExternalEntityParserCreate. A struct size mismatch between ELEMENT_TYPE members causes storeAtts to read the attIndex member past allocated memory boundaries, resulting in failure to normalize whitespace in non-CDATA attributes or a wild pointer dereference causing a segfault. This vulnerability was introduced by the fix for CVE-2026-66046.

🎖@cveNotify

🚨 CVE-2026-72854
msgpack_unpacker_expand_buffer in src/unpack.c, reached through the public msgpack_unpacker_reserve_buffer API, computes its new buffer size using an unchecked size_t addition of the requested size and the amount already used. The doubling loop guards its own multiplication against overflow, but the addition in the loop condition is unguarded, so a request near SIZE_MAX wraps: the loop condition is already satisfied, the allocation is performed at the small pre-wrap size, and the function returns true. The caller is told the requested capacity was reserved when it was not, so a subsequent write of the requested length overflows the heap buffer. The library's own example/lib_buffer_unpack.c demonstrates the reserve-then-write pattern, and its defensive assert comparing capacity against the request is compiled out under NDEBUG. msgpack-c's own decode entry points do not derive the reservation size from untrusted input, so reaching this requires an integration that passes an attacker-influenced length to the reservation API, such as a length-prefixed streaming transport.

🎖@cveNotify

🚨 CVE-2026-53804
OTRS Community Edition contains an authenticated OS command injection vulnerability in the PGP encryption module that allows administrators to execute arbitrary operating-system commands by supplying crafted values for the PGP binary path and command options. Administrator-supplied configuration values are concatenated without sanitization into a shell command, enabling arbitrary command execution as the web server process user during normal ticket operations after the malicious configuration is deployed.

🎖@cveNotify

🚨 CVE-2026-72818
The URLS regular expression in nltk/tokenize/casual.py, compiled into TweetTokenizer.WORD_RE and applied by TweetTokenizer.tokenize, contains a naked-domain branch whose domain-label prefix [a-z0-9]+(?:[.\-][a-z0-9]+)* is unbounded. Input consisting of many alternating label separators can be partitioned in exponentially many ways, and because the branch also requires a trailing top-level domain that such input never supplies, the engine explores those partitions before failing at each offset. A few kilobytes of input therefore consumes seconds to minutes of single-threaded CPU, and the HANG_RE substitution performed before matching does not collapse the pattern. TweetTokenizer is intended for tokenizing untrusted social-media text, so any service that applies it, or the module-level casual_tokenize, to submitted text can be stalled per request without authentication. Version 3.10.1 bounds the label repetition.

🎖@cveNotify

🚨 CVE-2026-72846
Lightdash stores the webhook URL supplied with a scheduled delivery and later posts to it from sendWebhook in packages/backend/src/clients/GoogleChat/GoogleChatClient.ts and in packages/backend/src/clients/MicrosoftTeams/MicrosoftTeamsClient.ts. In affected versions both call fetch on the stored URL directly. The validatePublicHttpUrl helper in packages/backend/src/utils/ssrfProtection.ts, used for MCP server URLs, is not applied on either path, and the webhook fields carry no server-side URL constraint. A user able to create or trigger a scheduled delivery can therefore direct the server to issue POST requests to private, loopback and link-local addresses, including cloud metadata endpoints, and can distinguish reachable internal services from unreachable ones through the resulting errors. The upstream response is never returned to the requester; on a failure status its body is written to the server log instead. Version 1.146.4 routes both clients through postSchedulerWebhook from packages/backend/src/utils/schedulerWebhookValidation rather than calling fetch directly.

🎖@cveNotify

🚨 CVE-2026-72848
SitemapLoader.parse_sitemap in langchain_community/document_loaders/sitemap.py applies the documented restrict_to_same_domain control only to leaf url entries. The loop over url elements filters cross-domain locations, but the loop over nested sitemap elements passes the child loc straight to self.scrape_all([loc.text], "xml"), which reaches WebBaseLoader.scrape_all and an aiohttp GET, with no domain comparison and no check for private, loopback or link-local destinations. An attacker who controls or influences an ingested sitemap can therefore point a nested sitemap entry at an internal address and make the server fetch it even when the deploying application set restrict_to_same_domain to True specifically to confine outbound requests. The fetched content is parsed and surfaces in the returned Documents, so internal responses are disclosed to the caller rather than merely requested.

🎖@cveNotify

🚨 CVE-2026-77767
Reconmap's API applies a fallback authorization policy in apps/api/app/Program.cs that requires an authenticated user holding the administrator role, so controllers without their own attribute reject anonymous callers. The report preview action in apps/api/app/Controllers/ReportsController.cs carries [AllowAnonymous] and therefore opts out of that policy. PreviewReport loads the Project row named by the id path segment, loads the linked Organisation through the project's ClientId, and renders both into default-report-template.html, which prints the project name and description together with the client organisation's name, address and URL. No authentication, project membership or role check is performed. Because the id is the auto-increment primary key of the project table, an unauthenticated remote caller can walk sequential ids to retrieve the engagement details and client organisation of every project on the instance, and the 404 returned for a missing id reveals which project ids exist. Reconmap stores penetration-testing engagements, so the disclosed descriptions and client records are sensitive by nature.

🎖@cveNotify

🚨 CVE-2026-77087
Paperclip before 0.3.1 in default local_trusted mode fails to validate Host headers, allowing attackers to execute arbitrary commands via DNS rebinding. An attacker can craft a malicious webpage that, when visited by a developer running Paperclip locally, uses DNS rebinding to make authenticated API requests and execute commands through the process adapter.

🎖@cveNotify

🚨 CVE-2026-77814
is_path_trusted in scripts/iib/api.py compares the requested path against each allowed parent directory with path.startswith(parent_path), without appending a path separator. A directory whose name merely begins with an allowed path therefore satisfies the comparison, so where /data/images is allowed a request for /data/images_private/secret.txt is treated as trusted and served by FileResponse, disclosing files the confinement was meant to exclude. Whether the check applies depends on get_enable_access_control in scripts/iib/tool.py: it returns true when IIB_ACCESS_CONTROL is set to enable, false when set to disable, and otherwise true when the host Stable Diffusion WebUI was started with share, ngrok, listen or server_name, falling back to false. Confinement is therefore active in the network-exposed WebUI deployments that rely on it, while a standalone run with no such option serves every readable file regardless of this flaw. The fix compares against parent_path joined with os.sep.

🎖@cveNotify