| CVE-2008-6994 | Stack-based buffer overflow in the SaveAs feature (SaveFileAsWithFilter function) in win_util.cc in Google Chrome 0.2.149.27 allows user-assisted remote attackers to execute arbitrary code via a web page with a long TITLE element, which triggers the overflow when the user saves the page and a long filename is generated. NOTE: it might be possible to exploit this issue via an HTTP response that includes a long filename in a Content-Disposition header. | high | |
| CVE-2014-7934 | Use-after-free vulnerability in the DOM implementation in Blink, as used in Google Chrome before 40.0.2214.91, allows remote attackers to cause a denial of service or possibly have unspecified other impact via vectors related to unexpected absence of document data structures. | critical | |
| CVE-2011-1296 | Google Chrome before 10.0.648.204 does not properly handle SVG text, which allows remote attackers to cause a denial of service or possibly have unspecified other impact via unknown vectors that lead to a "stale pointer." | critical | |
| CVE-2016-1617 | The CSPSource::schemeMatches function in WebKit/Source/core/frame/csp/CSPSource.cpp in the Content Security Policy (CSP) implementation in Blink, as used in Google Chrome before 48.0.2564.82, does not apply http policies to https URLs and does not apply ws policies to wss URLs, which makes it easier for remote attackers to determine whether a specific HSTS web site has been visited by reading a CSP report. | medium | |
| CVE-2016-1614 | The UnacceleratedImageBufferSurface class in WebKit/Source/platform/graphics/UnacceleratedImageBufferSurface.cpp in Blink, as used in Google Chrome before 48.0.2564.82, mishandles the initialization mode, which allows remote attackers to obtain sensitive information from process memory via a crafted web site. | medium | |
| CVE-2016-1618 | Blink, as used in Google Chrome before 48.0.2564.82, does not ensure that a proper cryptographicallyRandomValues random number generator is used, which makes it easier for remote attackers to defeat cryptographic protection mechanisms via unspecified vectors. | medium | |
| CVE-2026-54760 | Langroid is a framework for building large-language-model-powered applications. Prior to version 0.65.1, the `SQLChatAgent` SQL-injection mitigation, with default `allow_dangerous_operations=False`, combines a raw-text regex blocklist (`_DANGEROUS_SQL_PATTERNS`) with a `sqlglot` SELECT-only statement allowlist. The blocklist entries that target callable functions require the function name to be immediately followed by `\s*\(`. PostgreSQL accepts the same call with the name separated from `(` by a quoted identifier, an inline comment, or schema qualification. These forms evade the regex, still parse as `SELECT`, and execute the same PostgreSQL function. This restores the `pg_read_file` server-side file-read primitive that the prior CVE-2026-25879 / GHSA-pmch-g965-grmr fix was meant to block: the parent advisory fixed a missing `pg_read_file` blocklist entry, while this report shows that the added regex is bypassable. Version 0.65.1 fixes the issue. | critical | |
| CVE-2026-88874 | AVideo through revision c3edcc274c389816d434acadac07ee78eaf330c1 (master, 2026-08-23) does not enforce the Live stream password check on the stats endpoint or on the HLS origin. Live::_getStats() (plugin/Live/Live.php) returns a password-protected transmission's RTMP stream key, its isPasswordProtected flag, and its HLS (m3u8) URL to unauthenticated callers, in both the public applications list and the hidden_applications branch used when canSeeLiveFromLiveKey() fails. Separately, the shipped NGINX configuration (deploy/nginx/nginx.conf) serves the .m3u8 playlist, the AES-128 key, and the transport-stream segments from the /live location without any auth_request (the auth_key_check directive in the .key location is commented out). A remote, unauthenticated attacker can therefore retrieve the stream key and decryption key and watch a password-protected live transmission without supplying the configured password. No patched version was available at the time of the advisory. | high | |
| CVE-2026-108390 | Hestia Control Panel is a lightweight and powerful control panel for the modern web. From version 1.9.7 through 1.10.4, update_object_value() in func/main.sh does not escape single quotes when writing a custom document-root value to a user's web.conf file, allowing a low-privileged authenticated panel user with a web domain to inject an additional PATH field. When the user invokes the ordinary Search feature, the root-privileged v-search-user-object script loads the forged field through parse_object_kv_list_non_eval(), which permits reserved shell-variable names and globally replaces PATH before invoking column without an absolute path. By placing a malicious column executable in the injected search path, the user can execute arbitrary code as root. This residual vulnerability was introduced when version 1.9.7 replaced the raw eval sink addressed by CVE-2026-84984 with parse_object_kv_list_non_eval(), and it is fixed in version 1.10.5. | critical | |
| CVE-2026-11414 | A hard-coded cryptographic key is used by Altium Enterprise Server to sign file download URLs in the Vault service. Because the key is identical across all installations, an unauthenticated network attacker who can reach the server can forge valid download signatures and retrieve files from the Vault storage area without any authentication, session, or credentials. A separate path traversal vulnerability in the same download endpoint allows the configured storage root to be escaped, enabling reads of arbitrary files on the server filesystem. Combined, these issues allow an unauthenticated attacker to obtain sensitive server configuration and key material, which can lead to full server compromise. The vulnerability can be chained with CVE-2026-9152 to enumerate and bulk-download stored content. Altium 365 cloud deployments are not impacted in practice, as file storage uses object storage rather than the local filesystem. | critical | |
| CVE-2010-1822 | WebKit, as used in Apple Safari before 4.1.3 and 5.0.x before 5.0.3 and Google Chrome before 6.0.472.62, does not properly perform a cast of an unspecified variable, which allows remote attackers to execute arbitrary code or cause a denial of service (application crash) via an SVG element in a non-SVG document. | high | |
| CVE-2010-3114 | The text-editing implementation in Google Chrome before 5.0.375.127, and webkitgtk before 1.2.6, does not check a node type before performing a cast, which has unspecified impact and attack vectors related to (1) DeleteSelectionCommand.cpp, (2) InsertLineBreakCommand.cpp, or (3) InsertParagraphSeparatorCommand.cpp in WebCore/editing/. | high | |
| CVE-2010-4198 | WebKit, as used in Google Chrome before 7.0.517.44, webkitgtk before 1.2.6, and other products, does not properly handle large text areas, which allows remote attackers to cause a denial of service (memory corruption) or possibly have unspecified other impact via a crafted HTML document. | high | |
| CVE-2010-4204 | WebKit, as used in Google Chrome before 7.0.517.44, webkitgtk before 1.2.6, and other products, accesses a frame object after this object has been destroyed, which allows remote attackers to cause a denial of service or possibly have unspecified other impact via unknown vectors. | critical | |
| CVE-2010-3113 | Google Chrome before 5.0.375.127, and webkitgtk before 1.2.5, does not properly handle SVG documents, which allows remote attackers to cause a denial of service (memory corruption) or possibly have unspecified other impact via unknown vectors related to state changes when using DeleteButtonController. | critical | |
| CVE-2020-6519 | Policy bypass in CSP in Google Chrome prior to 84.0.4147.89 allowed a remote attacker to bypass content security policy via a crafted HTML page. | medium | |
| CVE-2014-1748 | The ScrollView::paint function in platform/scroll/ScrollView.cpp in Blink, as used in Google Chrome before 35.0.1916.114, allows remote attackers to spoof the UI by extending scrollbar painting into the parent frame. | high | |
| CVE-2016-1637 | The SkATan2_255 function in effects/gradients/SkSweepGradient.cpp in Skia, as used in Google Chrome before 49.0.2623.75, mishandles arctangent calculations, which allows remote attackers to obtain sensitive information via a crafted web site. | medium | |
| CVE-2016-1630 | The ContainerNode::parserRemoveChild function in WebKit/Source/core/dom/ContainerNode.cpp in Blink, as used in Google Chrome before 49.0.2623.75, mishandles widget updates, which makes it easier for remote attackers to bypass the Same Origin Policy via a crafted web site. | high | |
| CVE-2016-1636 | The PendingScript::notifyFinished function in WebKit/Source/core/dom/PendingScript.cpp in Google Chrome before 49.0.2623.75 relies on memory-cache information about integrity-check occurrences instead of integrity-check successes, which allows remote attackers to bypass the Subresource Integrity (aka SRI) protection mechanism by triggering two loads of the same resource. | critical | |
| CVE-2010-4203 | WebM libvpx (aka the VP8 Codec SDK) before 0.9.5, as used in Google Chrome before 7.0.517.44, allows remote attackers to cause a denial of service (memory corruption) or possibly execute arbitrary code via invalid frames. | critical | |
| CVE-2026-33472 | Cryptomator is an open-source client-side encryption application for cloud storage. Version 1.19.1 contains a logic flaw in CheckHostTrustController.getAuthority() that allows an attacker to bypass the security fix for CVE-2026-32303. The method hardcodes the URI scheme based on port number, causing HTTPS URLs with port 80 to produce the same authority string as HTTP URLs, which defeats both the consistency check and the HTTP block validation. An attacker with write access to a cloud-synced vault.cryptomator file can craft a Hub configuration where apiBaseUrl and authEndpoint use HTTPS with port 80 to pass auto-trust validation, while tokenEndpoint uses plaintext HTTP. The vault is auto-trusted without user prompt, and a network-positioned attacker can intercept the OAuth token exchange to access the Cryptomator Hub API as the victim. This issue has been fixed in version 1.19.2. | medium | |
| CVE-2026-48814 | Network-AI is a TypeScript/Node.js multi-agent orchestrator. In versions 5.7.1 and earlier, the MCP SSE server allows unauthenticated cross-origin MCP tool invocation due to an empty default secret. This issue was partially addressed by CVE-2026-46701 in version 5.4.5 by closing the CORS flaw (with Access-Control-Allow-Origin now set only for localhost origins), but the empty-default-secret flaw described in the title remained: the SSE MCP server still defaulted to an empty secret, _isAuthorized() still returned true when the secret was empty, and a non-loopback bind only produced a warning. As a result, the server still ran fully unauthenticated by default. Any non-browser caller (for example, curl, SSRF, or a 0.0.0.0 bind) could invoke all 22 MCP tools (config_set, agent_spawn, blackboard_write, token_*) with no credentials. This issue was fixed in version 5.7.2. | critical | |
| CVE-2026-77054 | boneIO Black stores passwords, including the MQTT broker and web panel passwords, in cleartext. The GET /api/config endpoint returns the full configuration with sub-services passwords exposed. Any authenticated user can call the endpoint and access the administrative passwords in cleartext form. Critically, due to CVE-2026-77053 (authentication disabled by default), an unauthenticated attacker on the same local network can retrieve all passwords. This issue was fixed in version 1.6.0: web panel passwords are stored as salted scrypt hashes in a file readable only by the service account, and the API and logs no longer return secret values. The MQTT broker password has to stay recoverable because the application presents it to the broker as a client; it is kept in a separate file readable only by the service account and is no longer exposed through the API. | medium | |
| CVE-2026-118517 | The Symfony HtmlSanitizer component UrlSanitizer parses URLs differently from browsers, allowing the allowLinkHosts() and allowMediaHosts() checks to be bypassed for URLs without a well-formed authority, resulting in untrusted hosts being referenced in sanitized output in Symfony versions 6.1.x prior to 6.4.50, 7.x prior to 7.4.23, and 8.x prior to 8.1.11. UrlSanitizer parses URLs with league/uri, which follows RFC 3986, while browsers follow the WHATWG URL standard. For URLs without a well-formed authority, such as http:evil.example, ///evil.example, or https://://evil.example/ the sanitizer finds no host, whereas browsers resolve them to evil.example. The guard added for CVE-2026-45066 missed the zero-slash form, and the host 0 was also skipped because PHP treats it as falsy. This is exploitable when a host allow-list is combined with allowRelativeLinks(true) or allowRelativeMedias(true). The default configuration is not affected. | low | |
| CVE-2026-40199 | Net::CIDR::Lite versions before 0.23 for Perl mishandles IPv4 mapped IPv6 addresses, which may allow IP ACL bypass. _pack_ipv6() includes the sentinel byte from _pack_ipv4() when building the packed representation of IPv4 mapped addresses like ::ffff:192.168.1.1. This produces an 18 byte value instead of 17 bytes, misaligning the IPv4 part of the address. The wrong length causes incorrect results in mask operations (bitwise AND truncates to the shorter operand) and in find() / bin_find() which use Perl string comparison (lt/gt). This can cause find() to incorrectly match or miss addresses. Example: my $cidr = Net::CIDR::Lite->new("::ffff:192.168.1.0/120"); $cidr->find("::ffff:192.168.2.0"); # incorrectly returns true This is triggered by valid RFC 4291 IPv4 mapped addresses (::ffff:x.x.x.x). See also CVE-2026-40198, a related issue in the same function affecting malformed IPv6 addresses. | medium | |
| CVE-2026-100602 | ClawHub (openclaw/clawhub application/backend) contains a missing authorization check in the changelog preview feature. A signed-in caller can invoke the public skills:generateChangelogPreview action for a skill they are not authorized to access; the previous version is read without the file-read authorization enforced on normal content access, and up to 8,000 characters of quarantined content may be submitted to the AI provider and reflected in the preview returned to the caller, disclosing restricted skill content. The issue was confirmed at revision cbfee7343ddc867316dd9b3de6fa8856730f9f41; the complete historical affected range was not established. It is fixed by PR #3682, included in revision 8c2de6c506bb4efabe3f0c2ffb8370b9e23d4650, which was deployed to clawhub.ai on 2026-09-11; self-hosted deployments should update to that revision or a later descendant. The npm CLI and OpenClaw runtime are separate products and are not affected. | high | |
| CVE-2026-75554 | Insufficient Session Expiration vulnerability in the OAuth token refresh grant in hexpm hexpm allows a user removed from an organization to keep reading its private packages. expand_repositories_scope/3 in lib/hexpm/permissions.ex only rewrites the literal repositories scope, so an explicitly granted repository:<org> or docs:<org> scope passes through it untouched. The refresh grant re-derives a new token from the stored granted_scopes, which holds that expanded form, so every refresh reproduces the organization scope without revisiting membership. Because both CDN edges authorize repository access from the token claim without querying the database, an account removed from an organization retains read access for as long as it keeps refreshing, bounded by the 30 day refresh token lifetime rather than the 30 minute access token lifetime. This issue affects hex.pm: from 2025-10-10 before 2026-08-24. | low | |
| CVE-2026-66073 | RabbitMQ is a messaging and streaming broker. From 3.13.0 until 3.13.15 and 4.0.20 and 4.1.11 and 4.2.6, Atom table exhaustion via management API node field. pUT /api/queues/:vhost/:name (and the exchanges and bindings endpoints) accepts a node JSON field. The value goes through rabbitnodes:make → listtoatom with no cluster membership check first. Each unique value permanently leaks one atom. A March 2026 refactoring (ea61ce2563) introduced safe helpers in rabbitmgmtnodes.erl (parsenodename, safeatom, and requirenodename, using binarytoexistingatom) and fixed several callers (QQ replica ops, wmauthattempts, wmnodememoryets, getsortreverse, and rabbitfederationmgmt), but getnode/1 in rabbitmgmtutil.erl:880-885, the primary vector used by directrequest/6, was not Roughly 900K requests crash the VM via systemlimit, and all tenants lose Any user with the management tag and one vhost, the lowest privilege This issue is fixed in versions 3.13.15 and 4.0.20 and 4.1.11 and 4.2.6. | medium | |
| CVE-2026-31786 | In the Linux kernel, the following vulnerability has been resolved: Buffer overflow in drivers/xen/sys-hypervisor.c The build id returned by HYPERVISOR_xen_version(XENVER_build_id) is neither NUL terminated nor a string. The first causes a buffer overflow as sprintf in buildid_show will read and copy till it finds a NUL. 00000000 f4 91 51 f4 dd 38 9e 9d 65 47 52 eb 10 71 db 50 |..Q..8..eGR..q.P| 00000010 b9 a8 01 42 6f 2e 32 |...Bo.2| 00000017 So use a memcpy instead of sprintf to have the correct value: 00000000 f4 91 51 f4 dd 00 9e 9d 65 47 52 eb 10 71 db 50 |..Q.....eGR..q.P| 00000010 b9 a8 01 42 |...B| 00000014 (the above have a hack to embed a zero inside and check it's returned correctly). This is XSA-485 / CVE-2026-31786 | high | |
| CVE-2026-30796 | Cleartext Transmission of Sensitive Information, Insufficiently Protected Credentials vulnerability in rustdesk-client RustDesk Client rustdesk-client on Windows, MacOS, Linux, iOS, Android (Address book sync, Heartbeat sync loop modules) allows Sniffing Attacks. The client places the preset address-book password verbatim into the heartbeat sync JSON body (src/hbbs_http/sync.rs). Over an intact HTTPS session it is not exposed in transit, but it is a reusable shared secret rather than a zero-knowledge proof, so it is recovered by any party that becomes the API endpoint - under the re-homed/rogue API server (CVE-2026-30797) - and the leaked credential then authorizes the server-side address book. This vulnerability is associated with program files src/hbbs_http/sync.rs and program routines heartbeat sync body builder (emits preset-address-book-password). This issue affects RustDesk Client: through 1.4.8. | medium | |
| CVE-2026-120754 | HDF5 through 2.2.0 contains an uninitialized memory read vulnerability in H5Z__filter_scaleoffset() in src/H5Zscaleoffset.c. The decode path validates reads against the allocated size of the chunk buffer rather than the number of bytes actually read from the file. Because H5D__chunk_lock() deliberately sizes filtered chunk buffers to the larger of the stored and unfiltered chunk sizes and fills only the stored bytes, a truncated chunk passes these bounds checks and is decompressed from the uninitialized remainder of the allocation. Unrelated heap contents are then returned to the caller as dataset data. Because the reads remain within the allocation, the issue is not detected by AddressSanitizer. An attacker who can get a user or application to read a dataset from a crafted HDF5 file can cause disclosure of process memory. This issue exists because of an incomplete fix for CVE-2026-19029. | medium | |
| CVE-2026-63301 | In Quick.CMS, the administrative user interface restricts deletion of the primary language by omitting the corresponding option from the interface; however, the underlying language-deletion API endpoint does not enforce an equivalent server-side authorization check. As a result, an authenticated administrator can bypass the UI-level restriction and delete the primary language by sending a direct HTTP request to the API endpoint. Successful deletion of the primary language results in a Denial of Service (DoS) of application. Critically, when combined with a separate Cross-Site Request Forgery (CSRF) vulnerability (CVE-2026-1468) an unauthenticated remote attacker can craft a malicious link, which if visited by an authenticated administrator, will trigger the DoS condition without direct access to the application The vendor assessed the likelihood of exploitation as very low and determined that a fix is not necessary. | high | |
| CVE-2022-31176 | Grafana Image Renderer is a Grafana backend plugin that handles rendering of panels & dashboards to PNGs using a headless browser (Chromium/Chrome). An internal security review identified an unauthorized file disclosure vulnerability. It is possible for a malicious user to retrieve unauthorized files under some network conditions or via a fake datasource (if user has admin permissions in Grafana). All Grafana installations should be upgraded to version 3.6.1 as soon as possible. As a workaround it is possible to [disable HTTP remote rendering](https://grafana.com/docs/grafana/latest/setup-grafana/configure-grafana/#plugingrafana-image-renderer). | high | |
| CVE-2011-2821 | Double free vulnerability in libxml2, as used in Google Chrome before 13.0.782.215, allows remote attackers to cause a denial of service or possibly have unspecified other impact via a crafted XPath expression. | high | |
| CVE-2010-4206 | Array index error in the FEBlend::apply function in WebCore/platform/graphics/filters/FEBlend.cpp in WebKit, as used in Google Chrome before 7.0.517.44, webkitgtk before 1.2.6, and other products, allows remote attackers to cause a denial of service and possibly execute arbitrary code via a crafted SVG document, related to effects in the application of filters. | high | |
| CVE-2011-1293 | Use-after-free vulnerability in the HTMLCollection implementation in Google Chrome before 10.0.648.204 allows remote attackers to cause a denial of service or possibly have unspecified other impact via unknown vectors. | critical | |
| CVE-2011-1659 | Integer overflow in posix/fnmatch.c in the GNU C Library (aka glibc or libc6) 2.13 and earlier allows context-dependent attackers to cause a denial of service (application crash) via a long UTF8 string that is used in an fnmatch call with a crafted pattern argument, a different vulnerability than CVE-2011-1071. | high | |
| CVE-2014-9862 | Integer signedness error in bspatch.c in bspatch in bsdiff, as used in Apple OS X before 10.11.6 and other products, allows remote attackers to execute arbitrary code or cause a denial of service (heap-based buffer overflow) via a crafted patch file. | high | |
| CVE-2019-13450 | In the Zoom Client through 4.4.4 and RingCentral 7.0.136380.0312 on macOS, remote attackers can force a user to join a video call with the video camera active. This occurs because any web site can interact with the Zoom web server on localhost port 19421 or 19424. NOTE: a machine remains vulnerable if the Zoom Client was installed in the past and then uninstalled. Blocking exploitation requires additional steps, such as the ZDisableVideo preference and/or killing the web server, deleting the ~/.zoomus directory, and creating a ~/.zoomus plain file. | medium | |
| CVE-2015-1209 | Use-after-free vulnerability in the VisibleSelection::nonBoundaryShadowTreeRootNode function in core/editing/VisibleSelection.cpp in the DOM implementation in Blink, as used in Google Chrome before 40.0.2214.111 on Windows, OS X, and Linux and before 40.0.2214.109 on Android, allows remote attackers to cause a denial of service or possibly have unspecified other impact via crafted JavaScript code that triggers improper handling of a shadow-root anchor. | high | |
| CVE-2023-51772 | One Identity Password Manager before 5.13.1 allows Kiosk Escape. This product enables users to reset their Active Directory passwords on the login screen of a Windows client. It launches a Chromium based browser in Kiosk mode to provide the reset functionality. The escape sequence is: wait for a session timeout, click on the Help icon, observe that there is a browser window for the One Identity website, navigate to any website that offers file upload, navigate to cmd.exe from the file explorer window, and launch cmd.exe as NT AUTHORITY\SYSTEM. | high | |
| CVE-2026-48824 | Mailpit is an email testing tool and API for developers. Prior to version 1.30.1, the fix for GHSA-fpxj-m5q8-fphw (CVE-2026-45710, "Mailpit: Set a default 50MB p/m limit to prevent DoS via unlimited SMTP DATA and /api/v1/send body sizes") wrapped only `POST /api/v1/send` with `http.MaxBytesReader`. The four other Mailpit JSON-body API endpoints `PUT /api/v1/messages` (SetReadStatus), `DELETE /api/v1/messages` (DeleteMessages), `PUT /api/v1/tags` (SetMessageTags), and `POST /api/v1/message/{id}/release` (ReleaseMessage) still call `json.NewDecoder(r.Body)` directly with no body-size cap and remain reachable unauthenticated in the default `docker run axllent/mailpit:latest` deploy. An unauthenticated remote attacker can post a multi-million-element `IDs` slice and drive RSS from ~25 MiB baseline to ~450 MiB per 16 MB request body. Repeating across multiple connections accumulates the same per-request amplification per process. Version 1.30.1 contains a patch. | medium | |
| CVE-2026-67260 | Apache Airflow 3.3.0 moved human-in-the-loop tasks from the triggerer to a new `awaiting_input` task state swept by the scheduler. That sweep deserializes the task instance's `next_kwargs` without an allow-list, so a Dag author — who controls that value through the task execution API — can cause an arbitrary module import and object instantiation inside the scheduler process, or terminate the scheduler job. No non-default configuration is required: the sweep runs unconditionally every 15 seconds, and the default `allowed_deserialization_classes` setting does not cover this code path. Versions before 3.3.0 are not affected, because human-in-the-loop tasks deferred onto the triggerer instead. This is a different code path from CVE-2026-58076, which covers the same unguarded exception-node deserialization reached elsewhere — deployments that applied that fix must upgrade for this issue as well. Users are advised to upgrade to apache-airflow 3.3.1 or later. | high | |
| CVE-2026-123012 | In Gogs 0.14.3, the Git Smart HTTP authorization layer in internal/route/repo/http.go allows a push when only read access was intended, a different vulnerability than CVE-2026-52810. This occurs because of a case-sensitivity desync between the authorization classifier and the request dispatcher: the gitHTTPIsPull function compares the Git action verb case-sensitively and defaults to pull (read access) for any unrecognized string, whereas the request dispatcher lowercases the same verb before routing it to the receive-pack handler. Capitalizing a single letter in the URL (e.g. info/Refs instead of info/refs, or git-receive-Pack instead of git-receive-pack) causes the classifier to misidentify a push as a pull, granting only read-level authorization, whereas the dispatcher routes it to the write handler. An authenticated read-only collaborator can exploit this to push to a private repository, creating, updating, or deleting refs. | high | |
| CVE-2026-54626 | SAIL is a cross-platform library for loading and saving images with support for animation, metadata, and ICC profiles. In 0.9.10 and earlier, the TGA_INDEXED_RLE path selected by image_type == 9 allocates an image buffer using the one-byte-per-pixel SAIL_PIXEL_FORMAT_BPP8_INDEXED format returned by tga_private_sail_pixel_format() in src/sail-codecs/tga/helpers.c, while sail_codec_load_frame_v8_tga() in src/sail-codecs/tga/tga.c derives a two-to-four-byte pixel_size from an attacker-controlled header bpp value from 9 through 32. Loading a crafted color-mapped run-length-encoded TGA through sail_load_from_file() or sail_load_from_memory() therefore writes attacker-controlled bytes beyond the heap pixel buffer. The pixel-count clamp added for CVE-2026-40494 does not constrain the per-pixel write width, so this issue is an incomplete fix of that vulnerability and can cause heap corruption, a reliable crash, or potential code execution. This issue is fixed in version 1.0.0. | critical | |
| CVE-2026-66065 | Ouroboros is a local-first runtime for AI coding agents that records their actions and applies user-defined policies to constrain behavior. Versions prior to 0.42.1 have an incomplete denylist. Several execution-routing keys of the same RCE class were omitted, so a malicious cloned repo can still reach arbitrary command execution by shipping a .env (auto-loaded at import, with no review step). The CVE-2026-47211 fix added _UNTRUSTED_ENV_DENYLIST to stop an untrusted project-directory .env from redirecting execution, but it did not account for all keys. The backend config-home and MCP/plugin roots bypass the approval gate by pointing the nested agent, MCP servers, and plugin roster at attacker config. Other variables re-enable blocked local transports, replace sub-agent prompts, switch backends, and lower tool approval classes, further weakening the approval gate. This issue has been fixed in version 0.42.1. | high | |
| CVE-2026-100601 | ClawHub (openclaw/clawhub) application/backend contains a server-side request forgery vulnerability in the public profile preview's image fetching. The preview accepts a user-supplied image URL and checks the textual hostname against private-address patterns, but does not validate or pin the resolved network destination, so a public-looking hostname can resolve to an internal address or change resolution between validation and connection (DNS rebinding). A maintainer-run local harness demonstrated an outbound connection to an owner-controlled loopback listener; access to production internal services, credential disclosure, and code execution were not demonstrated. The issue was confirmed at revision cbfee7343ddc867316dd9b3de6fa8856730f9f41; the complete historical affected range was not established. Fixed by PR #3683, included in revision 8c2de6c506bb4efabe3f0c2ffb8370b9e23d4650, which was deployed to clawhub.ai on 2026-09-11; self-hosted deployments should update to that revision or a later descendant. The npm CLI and OpenClaw runtime are separate products and are not affected. | medium | |
| CVE-2026-89793 | In the Linux kernel, the following vulnerability has been resolved: ublk: clear VM_MAYWRITE on read-only ublk char device mmap ublk_ch_mmap() rejects mmap requests with VM_WRITE set, but never clears VM_MAYWRITE on the resulting read-only mapping. This allows a userspace daemon to mmap the per-queue command buffer PROT_READ, then upgrade it to PROT_WRITE via mprotect(), since VM_MAYWRITE was never cleared. The command buffer holds struct ublksrv_io_desc entries that are kernel-written ABI; a writable mapping lets an unprivileged daemon process corrupt fields such as addr, op_flags, nr_sectors, and start_sector. Same bug class as the drm/panthor and drm/vc4 VM_MAYWRITE fixes, and the 2026-08-13 ptp/vmclock fix (a5edadbae57e). Verified via mprotect() PoC: before the fix, a PROT_READ mapping can be upgraded to PROT_READ|PROT_WRITE and a write into the command buffer corrupts io_desc fields (confirmed under KASAN). After the fix, mprotect() returns -EACCES. | high | |
| CVE-2010-4197 | Use-after-free vulnerability in WebKit, as used in Google Chrome before 7.0.517.44, webkitgtk before 1.2.6, and other products, allows remote attackers to cause a denial of service or possibly have unspecified other impact via vectors involving text editing. | critical | |
