CVE-2026-61979
Unauthenticated Privilege Escalation in SAML SP Single Sign On <= 5.4.3 versions.
CVE-2026-61946
Unauthenticated Insecure Direct Object References (IDOR) in Easy Appointments <= 3.12.27 versions.
CVE-2026-61876
LuCI versions fail to properly encode DHCPv6 lease hostnames before rendering in status tables, allowing adjacent network attackers to inject HTML markup. Attackers can send a DHCPv6 Client FQDN containing script tags that execute in the administrator's browser when viewing DHCP lease pages.
CVE-2026-61797
https://github.com/itres-labs/CVE-2026-61797
CVE-2026-61732
Decepticon is an autonomous hacking agent for red teams. Versions prior to 1.1.17 wrap web crawl results — the output of agent reconnaissance against target services — into LLM messages without neutralizing ChatML special-token literals. Under the BYOK (Bring Your Own Key) deployment model, users configure their own LLM credentials to any OpenAI-compatible endpoint. Most open-source and self-deployed model providers (vLLM, SGLang, Ollama, LM Studio, text-generation-webui, etc.) do not filter special-token literals from user content in their default configurations. Those literals are parsed into structural role-boundary token IDs, meaning an attacker string planted in a target web page forges a new operator turn the model treats as authoritative, bypassing Decepticon's agent guardrails and resulting in arbitrary command execution inside the Kali Linux sandbox. Version 1.1.17 patches the issue.
CVE-2026-61628
nginx ignition is a user interface for the nginx web server. Prior to version 2.41.1, `POST /api/users/onboarding/finish` is registered as anonymous (unauthenticated) and creates a user with full ReadWrite admin permissions. Because the handler uses a check-then-act (TOCTOU) pattern between the "onboarding already completed?" check and the user-creation write, with no atomic guard, a remote unauthenticated attacker who can reach an instance in its pre-onboarding state can create an administrator account for themselves — and concurrent requests can create multiple admin accounts in a single race. Version 2.41.1 patches the issue.
CVE-2026-61578
https://github.com/HORKimhab/CVE-Chamilo-LMS
CVE-2026-61515
Puwell IP Camera firmware versions 2.x through 4.x contains an unauthenticated command injection vulnerability that allows remote attackers to execute arbitrary operating system commands by sending a crafted JSON payload to the DebugShell interface exposed on TCP port 34567. Attackers can exploit the lack of authentication and input sanitization in the binary protocol service to pass arbitrary commands directly to the underlying operating system, achieving root-level code execution and complete device compromise.
CVE-2026-61511
vBulletin 5.x through 5.7.5 and 6.x through 6.2.1 contains an eval injection vulnerability in the vB5_Template_Runtime::runMaths() method within the template runtime that allows unauthenticated remote attackers to execute arbitrary PHP code by supplying crafted input through the pagenav[pagenumber] parameter. Attackers can exploit the insufficiently restrictive regex filter by using phpfuck-style encoding with permitted characters to inject and execute arbitrary PHP code via the unauthenticated ajax/render template route without any authentication.
CVE-2026-61500
Rejetto HFS 3.0.0 through 3.2.0 derives its session-cookie signing key from the non-cryptographic Math.random() generator and discloses outputs of the same generator to unauthenticated clients during login. A remote attacker can collect a small number of login responses, reconstruct the generator's state, recover the signing key, and forge a valid administrator session cookie, leading to full administrative access and remote code execution via the server_code configuration feature.
CVE-2026-61459
MCP Server Kubernetes before 3.9.0 contains an argument injection vulnerability in structured tools (kubectl_get, kubectl_describe, kubectl_delete) that allows attackers to bypass the assertNoDangerousFlags security check by supplying resourceType and name parameters with leading dashes. Attackers can inject the --server flag to redirect kubectl commands to an attacker-controlled API server, causing the operator's bearer token to be transmitted externally and enabling full cluster compromise.
CVE-2026-61447
PraisonAI before 1.6.78 contains a remote code execution vulnerability in CodeAgent._execute_python() that executes LLM-generated Python code without AST validation, import restrictions, or sandbox enforcement. Attackers can influence LLM output through prompt injection to exfiltrate all environment secrets and execute arbitrary code on the host system.
CVE-2026-61424
Joomla Extension - dj-extensions.com - Unauthenticated arbitrary file upload in DJ-Classifieds < 3.11.2 - The Joomla extension DJ-Classifieds is vulnerable to an unauthenticated file upload, leading to full RCE.
CVE-2026-61343
LibreBooking's email template editor save action passes the submitted template name directly into the destination file path, allowing a remote attacker with administrator credentials to write an arbitrary file outside the template directory and execute code. Fixed in 5.1.0.
CVE-2026-60206
Vulnerability in the Oracle WebLogic Server product of Oracle Fusion Middleware (component: Core). Supported versions that are affected are 12.2.1.4.0, 14.1.1.0.0, 14.1.2.0.0 and 15.1.1.0.0. Easily exploitable vulnerability allows low privileged attacker with network access via SAML to compromise Oracle WebLogic Server. While the vulnerability is in Oracle WebLogic Server, attacks may significantly impact additional products (scope change). Successful attacks of this vulnerability can result in takeover of Oracle WebLogic Server. CVSS 3.1 Base Score 9.9 (Confidentiality, Integrity and Availability impacts). CVSS Vector: (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:H).
CVE-2026-60137
WordPress 6.8.x before 6.8.6, 6.9.x before 6.9.5, and 7.0.x before 7.0.2 does not properly sanitise the author__not_in parameter of WP_Query, which could allow SQL Injection when a plugin or theme passes untrusted input to the parameter.
CVE-2026-60121
Vitec Flamingo 4.12.2 contains an unauthenticated OS command injection vulnerability in the admin/ajax/ping.php endpoint that allows remote attackers to execute arbitrary commands by exploiting a double-evaluation flaw in shell argument handling. The endpoint applies escapeshellarg() to the user-supplied host POST parameter before passing it to a system wrapper, but the wrapper retrieves the decoded value from argv and incorporates it into a second shell_exec() call without escaping, allowing injected commands to execute with root privileges via passwordless sudo.
CVE-2026-60093
Relative path traversal vulnerability in Apache Camel Azure-Storage Datalake component This issue affects Apache Camel: from 4.0.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. The camel-azure-storage-datalake component can download an Azure Data Lake Storage Gen2 file to the local filesystem through its downloadToFile operation, writing into the directory named by the fileDir endpoint option. DataLakeFileOperations.downloadToFile built the local target by joining fileDir with the remote path name exactly as the Azure SDK reported it (new File(fileDir, fileClientWrapper.getFileName())) and passed the result straight to the SDK download call, with no lexical normalization and no check that the resolved location stayed inside fileDir. The remote name is not route-controlled data: the consumer enumerates the filesystem in DataLakeConsumer.createBatchExchangesFromPath, which lists paths and creates one exchange per entry from PathItem.getName() verbatim, applying no name filtering by default. A path name containing parent-directory segments therefore resolved to a location outside the configured fileDir, letting anyone able to influence the names present in the consumed Data Lake filesystem cause Camel to create or overwrite a file at a location of their choosing, with the privileges of the Camel process. Depending on what the process can write to, overwriting a file outside the download directory can escalate beyond the loss of integrity of that file. The fileDir option is an ordinary common-group configuration parameter and carries no security marker, so nothing signalled to users that its value was not being enforced as a containment boundary. Camel's other file-download consumers - camel-file, camel-ftp, camel-smb, camel-mina-sftp and camel-azure-files - already constrained their local downloads to the configured directory using a path-segment boundary check; the camel-azure-storage-datalake download path was not covered by that work. Users are recommended to upgrade to version 4.22.0, which fixes the issue. If users are on the 4.14.x LTS releases stream, then they are suggested to upgrade to 4.14.9. If users are on the 4.18.x releases stream, then they are suggested to upgrade to 4.18.4. For deployments that cannot upgrade immediately, constrain the names the consumer will act on using the regex endpoint option, which is applied to each listed path name as a full-string match, so that only simple single-segment names are accepted and any name carrying a path separator or a parent-directory segment is filtered out before an exchange is created. Alternatively, avoid the downloadToFile operation on untrusted filesystems and write the payload from the route under a file name the route itself controls, rather than one taken from the remote listing. As defence in depth, treat the object names in any externally writable Data Lake filesystem as untrusted input and do not derive local filesystem paths from them.
CVE-2026-60004
Gitea before 1.27.1 allows remote code execution via the diffpatch API through Git hook installation.
CVE-2026-59970
Tandoor Recipes is an application for managing recipes, planning meals, and building shopping lists. Prior to 2.6.10, passes an administrator-controlled AI provider URL directly to LiteLLM without applying the server's outbound-request SSRF checks when an authenticated user creates a Space, configures an AI provider with a URL targeting an internal address, and invokes an AI action. In cookbook/views/api.py, the affected AI actions call completion through TARGET/api/ai-import/ and api/ai-step-sort/ while bypassing safe_request. The attack works because the server connects to the supplied provider URL from each AI action and exposes connection success, refusal, or timeout as a blind network-reachability oracle. This can reach internal services, other hosts on the server network, and cloud metadata endpoints and map reachable internal hosts and ports. This issue is fixed in version 2.6.10.
CVE-2026-59941
Dompdf is an HTML to PDF converter for PHP. Versions 3.15 and prior accept a BMP image and generates a PDF-compatible PNG based only on its declared header dimensions and never bounds width × height before the image is converted through GD. A 58-byte BMP whose header declares e.g. 6000×6000 is accepted and later drives imagecreatetruecolor($width, $height) (and PHP's native BMP decoder) to allocate the full pixel canvas. A payload can fit in a single HTTP request: the BMP can be inlined as a data:image/bmp;base64,… URI inside attacker-controlled HTML, so no upload, no remote fetch, and no chroot-reachable file is required. I measured a 169-byte request driving a dompdf render to ~412 MB peak RSS and ~4.8 s of CPU/wall time, versus ~34 MB for an identically-sized benign request — roughly a 12× memory amplification per request, repeatable and unauthenticated. This issue has been fixed in version 3.16.
CVE-2026-59903
Netty is an asynchronous, event-driven network application framework. Prior to 4.1.137.Final and 4.2.17.Final, io.netty.handler.codec.http.cors.CorsHandler setVaryHeader replaces application Vary headers such as Authorization or Cookie with Origin, allowing a caching proxy or CDN to reuse authenticated responses across users and disclose sensitive information. This issue is fixed in versions 4.1.137.Final and 4.2.17.Final.
CVE-2026-59891
sigstore-js provides JavaScript libraries for interacting with Sigstore services. Prior to 0.7.1, getRegistryCredentials() reads credentials from the Docker config file and selects an entry by checking whether any configured auth key contains the target registry string. Because this is a substring match rather than an exact host match, credentials configured for one registry can be selected for and transmitted to a different registry whose hostname has a substring relationship with a configured auth key. This issue is fixed in version 0.7.1.
CVE-2026-59827
Metabase is an open-source business intelligence and embedded analytics tool. Prior to 1.58.15, 1.59.12, 1.60.6.3, and 1.61.1.4, Metabase instances with an H2 database connection, including the default sample database, deserialize arbitrary Java objects returned in H2 native query result columns of type OTHER without validation, allowing an authenticated user who can run native H2 queries to execute code on the Metabase server. This issue is fixed in versions 1.58.15, 1.59.12, 1.60.6.3, and 1.61.1.4.
CVE-2026-59822
LiteLLM is a proxy server (AI Gateway) to call LLM APIs in OpenAI (or native) format. Prior to 1.84.0, LiteLLM's MCP Streamable HTTP endpoint allowed an unauthenticated attacker to use a fabricated Authorization header to trigger an OAuth2 passthrough fallback path that replaced failed LiteLLM key validation with an empty UserAPIKeyAuth() object, allowing requests to reach MCP tooling without a valid LiteLLM key. This issue is fixed in version 1.84.0.
CVE-2026-59774
Gitea's Org-mode markup renderer processes INCLUDE directives supplied to the repository markup endpoint without preventing access to server-side files. An unauthenticated attacker can use a suitable public repository to read files accessible to the Gitea service account. If the disclosed files include app.ini and its INTERNAL_TOKEN, the attacker can chain the disclosure with internal logger access to modify a Git hook and execute commands as the Gitea service account.
CVE-2026-59734
Coolify is an open-source and self-hostable tool for managing servers, applications, and databases. Prior to 4.0.0-beta.469, Coolify's app/Jobs/ApplicationDeploymentJob.php generate_healthcheck_commands() function directly interpolated the health_check_host, health_check_method, and health_check_path parameters into shell commands without proper sanitization, allowing authenticated users to execute arbitrary commands inside deployment containers. This issue is fixed in version 4.0.0-beta.469.
CVE-2026-59726
Ruflo is an agent meta-harness for Claude Code and Codex. Prior to 3.16.3, ruflo's default docker-compose deployment exposed the MCP bridge POST /mcp and POST /mcp/:group endpoints without authentication, allowing an unauthenticated network attacker to invoke tools/call to terminal_execute, obtain a shell in the bridge container, read provider API keys, and poison AgentDB learning-store patterns. This issue is fixed in version 3.16.3.
CVE-2026-59550
Unauthenticated SQL Injection in AWP Classifieds <= 4.4.7 versions.
CVE-2026-59358
Improper authentication (CWE-287) in the OAuth token endpoint in Cloud Foundry UAA allows a remote, authenticated attacker holding a valid user access token to obtain a fully-privileged client_credentials token for the OAuth client that issued it, by presenting the user token as an OAuth 2.0 Bearer credential on a client_credentials grant request in place of the client’s configured secret. UAA’s client_credentials handling does not verify that the Bearer credential supplied for client authentication is actually a client credential (a client secret or a valid configured client authentication method); it accepts any valid access token whose client_id matches the request. A token obtained by a normal end user through a public authorization_code + PKCE flow — scoped only to uaa.user, carrying a user_id, and recording client_auth_method=none — satisfies this check. That user token cannot itself administer OAuth clients (POST /oauth/clients correctly returns 403), but when replayed as Bearer authentication on a client_credentials request for the same client, UAA issues a new client-only token carrying the client’s full authorities, such as clients.write. An attacker can use that token to create arbitrary new OAuth clients, including clients with attacker-chosen authorities, without ever possessing the client’s actual secret. Exploitation requires a valid user access token (the attacker’s own) for a client that is configured to support both a public, user-facing authorization flow and the client_credentials grant type on the same client_id — a non-default combination. Practical impact scales with the authorities assigned to that client.
CVE-2026-59346
VMware Workstation and Fusion contain an integer-overflow vulnerability. A malicious actor with local administrative privileges on a virtual machine with VMXNET3 virtual network adapter may exploit this issue to execute code on the host. Affected versions: - VMware Workstation: 25H2, 26H1 (fixed in 26H1u1) - VMware Fusion: 25H2, 26H1 (fixed in 26H1u1)
CVE-2026-59310
VMware vCenter contains a directory traversal vulnerability in the Syslog server. A malicious actor with network access to vCenter may exploit this issue to execute arbitrary code.
CVE-2026-59265
A code execution issue in the Java integration in Apache OpenOffice v4.1.16 and earlier allows a crafted untrusted document to trigger executing arbitrary (even remote) code when opened by the user. This issue is expected to be fixed in version 4.1.17, which is in the release candidate phase. Until then, users can mitigate this issue by disabling Java runtime integration in the Preferences dialog. This prevents the attack. If this is not possible, or as an extra precaution, you can avoid opening open untrusted files entirely. Once 4.1.17 is released, upgrade to that version to fix the issue.
CVE-2026-59248
Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service. The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies. Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion. Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable. This issue affects cowlib: from 2.0.0 before 2.19.0.
CVE-2026-59243
The FAB auth manager's Azure AD OAuth login defaulted `verify_signature=False` when decoding the ID token, so an attacker able to present a forged or unsigned (`alg:none`) ID token to the OAuth callback could bypass authentication and log in as an arbitrary user, including one holding the Admin role (CWE-347). Deployments running the FAB auth manager with the Azure AD OAuth login path under its default configuration are affected; the Authentik path already defaulted to `True`. This issue affects `apache-airflow-providers-fab` before 3.7.3. Users are advised to upgrade to `apache-airflow-providers-fab` 3.7.3, which defaults `verify_signature=True`.
CVE-2026-59230
Improper input validation vulnerability in Apache Camel. This issue affects Apache Camel: from 2.17.0 before 4.14.9, from 4.15.0 before 4.18.4, from 4.19.0 before 4.22.0. The camel-mail component ships a MimeMultipart data format that can unmarshal a MIME multipart message. When it is configured with headersInline set to true, the unmarshal path copies the MIME headers of the incoming message onto the Camel message: it enumerates every header that is not one of the three standard ones it generates itself - Message-ID, MIME-Version and Content-Type - and calls setHeader for each, applying no HeaderFilterStrategy. The names of those MIME headers come from the message being unmarshalled, so a sender able to influence the message could place a header whose name falls in the Camel-internal namespace and have it set on the Exchange. Camel components read control headers from that namespace to override their configured behaviour - the camel-sql producer, for instance, takes the statement to execute from a Camel header when one is present - so an injected header could redirect what a downstream step in the route does with data the route author never intended it to take from the message. Which sinks are reachable, and what the consequences are, depends entirely on what the route does after the unmarshal step. The camel-mail consumer already applied a header filter strategy on its own inbound path, so this was the parallel inbound path into the same component that the earlier hardening did not cover. The affected copy is reached only when headersInline is enabled, which is not the default: with the default setting the MIME headers are surfaced as attachments rather than as message headers, and are not affected. The behaviour dates back to the introduction of the data format in 2.17.0 and was present on every release line until this fix. Users are recommended to upgrade to version 4.22.0, which fixes the issue. If users are on the 4.14.x LTS releases stream, then they are suggested to upgrade to 4.14.9. If users are on the 4.18.x releases stream, then they are suggested to upgrade to 4.18.4. For deployments that cannot upgrade immediately, leave headersInline at its default of false where the inline headers are not needed, since the copy is only reached when it is enabled. Where it must stay enabled, strip Camel-internal headers immediately after the unmarshal step, for example with removeHeaders(“Camel*”) placed before any processor or producer that reads control headers, and do not unmarshal MIME content from an untrusted sender into a route that dispatches on header values. As defence in depth, treat the header names of any MIME message arriving from outside the trust boundary as untrusted input.
CVE-2026-58635
Improper neutralization of special elements used in a command ('command injection') in Windows Narrator Braille allows an authorized attacker to elevate privileges locally.
CVE-2026-58480
Blocksy Companion Pro plugin for WordPress before 2.1.47 contains an unauthenticated arbitrary file upload vulnerability that allows attackers to upload executable files by bypassing extension validation in the save_attachments function exposed through the Advanced Reviews feature. Attackers can exploit the Custom Fonts extension's flawed strpos() substring check by uploading double-extension filenames such as shell.woff2.php, causing the validation to pass on the substring match while the web server executes the file as PHP, achieving remote code execution.
CVE-2026-58457
Shenzhen Aitemi M300 Wi-Fi Repeater (hardware model MT02) contains an unauthenticated OS command injection vulnerability that allows network-adjacent attackers to execute arbitrary shell commands by injecting unsanitized input through the smacfilter_conf handler in the commuos web backend. Attackers can append semicolon-delimited payloads to the name, enable, or mac GET parameters, which are passed without sanitization into sprintf() to build uci shell commands executed via doSystemCmdComlib(), granting full root-level control of the device.
CVE-2026-58455
Dockwatch through 0.6.567 contains an unauthenticated OS command injection vulnerability that allows remote attackers to execute arbitrary shell commands by exploiting a missing exit() after an authentication redirect in loader.php combined with unsanitized input passed to shell_exec() in ajax/compose.php. Attackers can seed the required session flag through the incomplete auth check, then inject arbitrary commands via the composePath POST parameter in the composePull action to achieve full host compromise, facilitated by the standard deployment mounting of the Docker socket.
CVE-2026-58289
Access of resource using incompatible type ('type confusion') in Microsoft Edge (Chromium-based) allows an unauthorized attacker to execute code over a network.
CVE-2026-58249
SAP NetWeaver and SAP Web Dispatcher do not properly validate certain fields in specially crafted requests handled by an internal protocol component. An unauthenticated attacker could send a malicious request that triggers a memory-handling error, causing the affected process to crash. Successful exploitation can therefore result in a high impact on the availability of the affected application. There is no impact on the confidentiality or integrity of the application.
CVE-2026-58231
SAP Commerce Cloud allows an unauthenticated attacker to abuse a default authentication client and submit specially crafted input to certain functions lacking sufficient validation. Successful exploitation could enable arbitrary code execution and compromise internal components, resulting in high impact on confidentiality, integrity, and availability of the application.
CVE-2026-58225
SQL Injection vulnerability in elixir-ecto postgrex allows an attacker who can influence a LISTEN channel name to inject SQL into the reconnect replay query, causing a denial of service of the notification connection. Postgrex.Notifications sanitizes channel names with quote_channel/1, which doubles double quotes so the name is safe inside a double-quoted identifier. This protects the single-statement LISTEN and UNLISTEN paths. On every (re)connect, however, handle_connect/1 replays all registered channels at once by concatenating their LISTEN statements and wrapping them in a dollar-quoted anonymous code block (DO $$BEGIN ... END$$). quote_channel/1 does not escape the $$ dollar-quote delimiter that opens and closes this block. The listen/3 guards only reject null bytes and names longer than 63 bytes, so a channel name containing $$ passes validation unchanged. Once such a name is embedded, its $$ prematurely terminates the outer dollar-quoted string and PostgreSQL parses the remainder as additional top-level statements. Because handle_connect/1 runs on every (re)connect, the malformed replay query is rejected each time and the notification connection never re-establishes its subscriptions, silently dropping notifications for every channel sharing that connection. An application is affected when it passes untrusted input (for example a tenant or user identifier) as a channel name to Postgrex.Notifications.listen/3. The double-quote doubling prevents forming a fully valid injected statement, so arbitrary SQL execution is not possible, but the corrupted query reliably breaks the shared notification connection for all tenants, resulting in denial of service. This issue affects postgrex: from 0.16.0 before 0.22.3.
CVE-2026-58222
A security flaw combining LDAP filter injection and improper authorization checks was found in Samba Active Directory Domain Controller (AD DC). When processing LDAP Compare requests, Samba fails to properly validate user-supplied attribute names and executes the resulting internal database search in a trusted context, bypassing normal Access Control List (ACL) enforcement. An authenticated low-privilege domain user can exploit these flaws to disclose confidential Active Directory attributes that would normally be inaccessible. The disclosed information may be leveraged to derive sensitive authentication material, potentially leading to privilege escalation and complete domain compromise. For example: In deployments configured with Group Managed Service Accounts (gMSAs), an attacker can extract the "msKds-RootKeyData" attribute and derive gMSA passwords offline, potentially leading to complete domain compromise if privileged gMSAs are present.
CVE-2026-58221
A flaw was found in Samba Active Directory Domain Controller (AD DC). Improper authorization checks allow an authenticated low-privilege domain user to modify internal LDB special records through LDAP. By altering the DSDB module configuration, an attacker can bypass directory ACL enforcement on new LDAP connections and elevate privileges, potentially leading to complete domain compromise.
CVE-2026-58138
Orkes Conductor 3.21.21 before 3.30.2 contains an unauthenticated remote code execution vulnerability that allows remote attackers to execute arbitrary OS commands by submitting inline workflow definitions containing malicious JavaScript or Python expressions to the workflow API endpoint prior to authentication. Attackers can exploit unsandboxed GraalVM evaluators configured with HostAccess.ALL or allowAllAccess(true) through INLINE, LAMBDA, DO_WHILE, and SWITCH task types to invoke arbitrary system commands via Java reflection or direct subprocess calls.
CVE-2026-58116
LLaMA-Factory through 0.9.5 contains a remote code execution vulnerability that allows attackers with WebUI access to execute arbitrary Python code by supplying a malicious model path in the Chat or Training interfaces. The application passes user-supplied model path input unvalidated into AutoTokenizer.from_pretrained() and AutoModel.from_pretrained() with a hardcoded trust_remote_code=True parameter, causing the Hugging Face transformers library to fetch and execute arbitrary code from a remote or local model repository with the privileges of the server process.
CVE-2026-58073
A vulnerability in Veeam Service Provider Console allowing an unauthenticated attacker to impersonate a managed agent andobtain that agent's credentials.
