Vulnerabilidades exploráveis hoje
17,404na visualização atual
Score único que combina CVSS, presença no KEV e EPSS. Cada CVE com sua própria ficha — timeline da publicação até a exploração ativa.
No catálogo KEV1,768
Novos KEV · 24H0
Exploit Today ≥ 701,700
Distribuição · última janela
- Crítica5,746
- Alta22,381
- Média21,187
- Baixa2,174
Filtros
Janela
Severidade
Filtros
CVECVSSEPSSKEVRExploitTítuloMod.
CVE-2026-973715.3 MÉD—
——0ntopng is a web-based network traffic monitoring application. Prior to 6.7.260721, ndpi_get_custom_category_match writes a four-byte nDPI protocol-category value through a pointer to the two-byte id variable in IpAddress::reloadBlacklist in src/IpAddress.cpp during packet dissection and host initialization. A remote unauthenticated attacker on a monitored network segment can send packets that match custom-category rules, causing stack corruption and denial of service without user interaction. This issue is fixed in version 6.7.260721.7hCVE-2026-852876.5 MÉD—
——0HomeBox is a home inventory and organization system. Prior to 0.27.0, POST /api/v1/users/register and the users.email uniqueness constraint treat email addresses as case-sensitive while HomeBox uses a case-insensitive login lookup. An unauthenticated attacker who knows a victim's email can register a differently cased variant, creating duplicate rows that cause login to return 401 for both variants and deny the victim access, while also bypassing the one-account-per-email policy. This issue is fixed in version 0.27.0.7hCVE-2026-829657.5 ALT—
——0A flaw was found in the SAML implementation of Keycloak, an open-source identity and access management solution. The SAML Artifact Resolution and ECP SOAP endpoints do not properly limit the size or complexity of incoming XML requests. An unauthenticated remote attacker can exploit this by sending specially crafted, large XML messages that cause the server to consume excessive memory. This leads to the application running out of memory and crashing, resulting in a denial of service for all users.5hCVE-2026-824138.8 ALT—
——0ntopng is a web-based network traffic monitoring application. Prior to 6.7.260717, scripts/lua/rest/v2/add/host/to_scan.lua and scripts/lua/rest/v2/exec/host/schedule_vulnerability_scan.lua accept scan_ports from any authenticated session, including non-administrator users, while deployments with login disabled expose the same request without authentication. The endpoints validate scan_ports with validateSingleWord, which permits shell metacharacters, and persist the value for the minute vulnerability-scan scheduler. The scheduler concatenates the value into an nmap command through nmap_scan_host and executes it through ntop.execCmd or ntop.execCmdAsync using popen. For the tcp_portscan and cve modules, a preliminary nmap -sn probe must first report the target as up before the attacker-controlled ports are used. When nmap is installed, this command injection allows arbitrary operating-system commands to execute as the ntopng service account without further user interaction. This issue is fixed in version 6.7.260717.7hCVE-2026-760778.1 ALT—
——0ASN1SCC is an open source ASN.1 compiler for embedded systems. Prior to 4.7.0.1, the XER tokenizer NT() in asn1crt/asn1crt_encoding_xer.c, which is reached by all XER decode paths, copies tag characters into a fixed char Value[100] field without a length bound and reads input and lookahead bytes without checking the buffer count. A crafted XER message with an overlong tag can overwrite the tokenizer's stack frame, while a truncated tag or unterminated processing-instruction or comment sequence can read beyond the input and may cause a CPU spin. The flaws can disclose adjacent memory, corrupt stack data, crash or hang the decoder, and potentially permit code execution. This issue is fixed in version 4.7.0.1.7hCVE-2026-756155.3 MÉD—
——0ASN1SCC is an open source ASN.1 compiler for embedded systems. Prior to 4.7.0.1, the ACN runtime function Acn_Dec_String_CharIndex_private in asn1crt/asn1crt_encoding_acn.c and the generated uPER decoder InternalItem_string_with_alpha_decode in StgC/uper_c.stg do not validate a decoded character index before reading allowedCharSet. For string alphabets whose size is not a power of two, a crafted message can select an unused encoded index, read a small bounded window of adjacent static program constants, and return the byte as a successfully decoded character. The full 128-character IA5String alphabet is not affected. This issue is fixed in version 4.7.0.1.7hCVE-2026-756148.1 ALT—
——0ASN1SCC is an open source ASN.1 compiler for embedded systems. Prior to 4.7.0.1, BitStream_DecodeConstraintWholeNumber in asn1crt/asn1crt_encoding.c reads enough bits for a constrained integer but does not reject values above the declared range when the encoded range is not a power of two. Generated uPER and ACN decoders for variable-size SEQUENCE OF, OCTET STRING, BIT STRING, and string values use the invalid result as a loop or copy length, so a crafted over-range size field can write beyond a destination buffer. Decoding untrusted affected data can corrupt heap or stack memory, cause a crash, or potentially permit code execution. This issue is fixed in version 4.7.0.1.5hCVE-2026-755995.4 MÉD—
——0HomeBox is a home inventory and organization system. Prior to 0.27.0, the maybeUrl() utility accepts markdown-style links in entity custom-field values without restricting the URL scheme, so an authenticated user can store a javascript: URI that is rendered as a clickable anchor on an item page. When another authenticated user clicks the link, attacker-controlled JavaScript executes in the victim's HomeBox origin, enabling session theft, account takeover, or data exfiltration. This issue is fixed in version 0.27.0.7hCVE-2026-541468.8 ALT—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. From 0.53.15 before 0.57.5 on macOS, core_process.rs in the Tauri desktop process publishes the per-launch OPENHUMAN_CORE_TOKEN bearer credential in its process environment even though the core now runs as an in-process task. Another process running as the same user can read that environment through the macOS KERN_PROCARGS2 process-inspection interface, recover the token together with the local host and port, and authenticate to the core JSON-RPC service with the authority of the desktop frontend. The attacker can read agent memory and integrated-account data, invoke privileged actions, change configuration, and stop or disrupt services without root access, developer mode, or changes to System Integrity Protection. This issue affects Intel and Apple Silicon macOS systems. This issue is fixed in version 0.57.5.7hCVE-2026-541458.8 ALT—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. In affected versions before 0.57.5, the approval gate uses the JSON-RPC bearer token as an approval session identifier. This exposes the credential through persisted approval records, events, debug logs, and RPC responses, including during normal desktop operation. An attacker who obtains these artifacts can recover the token and access the core RPC interface while it remains valid. This can expose sensitive agent and integration data and permit privileged operations. Stable tokens remain usable until rotated; launch-specific tokens remain usable while the corresponding core instance is active. The fix is included in version 0.57.5.7hCVE-2026-541446.1 MÉD—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0 on macOS, the runtime injected into a supported remote account webview such as Google Meet allows page JavaScript to call getDisplayMedia without a live user gesture. Instead of rejecting the request, the shim falls through to native Chromium capture behavior, and the embedded CEF environment can auto-select the primary display because it lacks the normal desktop-media picker. If OpenHuman already has macOS Screen Recording permission and an attacker obtains JavaScript execution in an allowed embedded origin, the attacker can obtain a real desktop capture stream without a new user gesture. The stream can expose chats, documents, source code, email, meeting content, credentials, one-time codes, and other visible desktop information. The separate direct screen_share_begin_session IPC path failing origin validation does not prevent exploitation through getDisplayMedia. This issue is fixed in version 0.56.0.7hCVE-2026-541439.8 CRÍ—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before version 0.57.13, an attacker who can send messages to a configured external channel can cause the agent to create persistent shell cron jobs without human approval. The scheduler later executes these jobs with the privileges of the OpenHuman runtime user. Command validation also permits leading environment assignments that can redirect allowed commands to executable helpers. The demonstrated exploit requires a suitable helper executable already present on the host. This can lead to unauthorized persistent command execution and compromise data and resources accessible to OpenHuman. This issue is fixed in version 0.57.13.7hCVE-2026-541428.7 ALT—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, background synchronization for connected services such as email, Slack, and Notion stores raw external content and later places it in the model context of autonomous background agents without neutralizing embedded instructions. ApprovalGate treats tool calls without an interactive chat context as pre-authorized automation, so attacker-authored content processed during a background tick can influence the model to invoke Composio actions without presenting an approval request. An attacker who can deliver content to a connected account can use this path to send or forward data, post messages, create or modify third-party content, or disclose information from OpenHuman memory and integrations. The issue requires an active integration and successful model interpretation of the injected instructions. This issue is fixed in version 0.57.13.5hCVE-2026-541419.0 CRÍ—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, a remote direct message received through a configured Yuanbao channel is converted into a generic channel AgentTurnRequest that uses the default tool policy without establishing an interactive approval context. Under the default Supervised autonomy policy with Node enabled, node_exec is classified as a write operation that should prompt for approval, but ApprovalGate treats missing APPROVAL_CHAT_CONTEXT as a non-interactive turn and allows execution. Because Yuanbao direct messages are open by default, an external sender who can message the connected bot can influence the agent to execute JavaScript and operating-system commands as the OpenHuman user without an approval prompt. The issue requires a configured Yuanbao channel, the default approval-requiring policy state, Node availability, and successful model selection of node_exec. This issue is fixed in version 0.57.13.7hCVE-2026-541408.3 ALT—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, OpenHuman places names, descriptions, titles, and parameter definitions returned by configured external MCP servers into the agent model context without neutralizing instructions or bounding the descriptive fields. The McpRemoteTool records produced by list_tools preserve this remote content, while tool filtering considers the tool name and the prompt-injection detector screens user messages rather than MCP definitions. A malicious or compromised MCP server can therefore return a poisoned description that is evaluated on every agent turn and may direct the model to invoke other available tools, disclose accessible data, or perform unauthorized actions without a separate malicious user message. The issue affects users who configure an attacker-controlled or compromised MCP server. This issue is fixed in version 0.57.13.7hCVE-2026-494077.7 ALT—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before version 0.57.44, its local /ws/dictation WebSocket endpoint lacks client authentication and browser-origin validation. An unauthenticated client, including a malicious web page able to reach the endpoint, can submit audio and trigger resource-intensive speech-to-text processing. Earlier affected versions also accumulate audio in an unbounded buffer, allowing continuous streaming to exhaust process memory. These conditions can degrade performance, exhaust resources, or crash the application. This issue is fixed in version 0.57.44.7hCVE-2026-494056.2 MÉD—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0, the local server exposes the server-sent events endpoint at /events without bearer authentication, permits cross-origin browser access, and uses only a caller-supplied client_id to select WebChannelEvent objects from the process-wide event bus. An unauthenticated local client, or an attacker-hosted web page opened by the victim, can subscribe to another client's stream by supplying a known or guessed identifier; the fixed system identifier makes the proactive-message stream predictable. The exposed events can include complete assistant responses, streamed message content, tool names, tool arguments, and raw tool outputs intended for the local user interface. This issue is fixed in version 0.56.0.5hCVE-2026-493938.4 ALT—
——0OpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.56.0, the core server protects HTTP POST requests to /rpc with bearer middleware but exposes Socket.IO rpc:request and chat:start handlers without equivalent transport or handler-level authentication. A local unauthenticated client can connect to the Socket.IO namespace and dispatch requests through the same protected controller path that rejects a missing bearer on /rpc. Depending on the selected method, the attacker can read configuration or health data, start agent flows, interact with integrated account APIs, alter local data, or stop and restart services. This issue is fixed in version 0.56.0.7hCVE-2026-123151——
——0A flaw was found in the GNU debugger (GDB). When GDB connects to a remote debug target, a malicious target can cause GDB to overflow a buffer on the stack while parsing the target's replies. This can crash GDB or potentially allow the target to execute arbitrary code in the GDB client process with the privileges of the user running GDB.7hCVE-2026-1219978.1 ALT—
——0A heap buffer overflow in gdbserver's QTDPsrc packet handler allows an unauthenticated remote attacker to achieve native code execution in the gdbserver process. The vulnerability results from a mismatch between a packet's declared source length and the actual data length, causing hex2bin to write beyond the allocated heap buffer.5hCVE-2026-1219968.2 ALT—
——0A length validation flaw in gdbserver's agent-expression parser allows an unauthenticated remote attacker to read uninitialized heap memory from the gdbserver process and cause a denial of service. Disclosed heap content includes addresses suitable for defeating ASLR in further exploitation.5hCVE-2026-1219957.5 ALT—
——0A stack buffer overflow in the GDB client's tracepoint upload handler allows an attacker-controlled debug stub to corrupt stack memory automatically during connection setup. The overflow results from GDB reading beyond the current reply boundary into a stale prior reply in the shared packet buffer.7hCVE-2026-1219948.8 ALT—
——0A stack buffer overflow in the GDB client's remote protocol handler allows an attacker-controlled debug target to execute arbitrary code in the GDB process. The vulnerability exists in fetch_register_using_p, which decodes a stub's register reply into an alloca-sized buffer without bounding writes by the register width.5hCVE-2026-1204997.8 ALT—
——0h5io before 0.2.6 contains a code injection vulnerability in read_hdf5() that allows attackers to execute arbitrary code by supplying a crafted HDF5 file. Attackers can store Python code in a node titled timezone, which _triage_read() passes to eval() with builtins reachable, name an arbitrary module in class or custom-object nodes that _import_class() imports, or place pickled objects in pd_dataframe or pd_series nodes that pandas deserializes on read.5hCVE-2026-1190837.8 ALT—
——0Path traversal in the archive extraction utility in GluonTS before 0.17.0 might allow an actor who supplies a crafted tar archive to write arbitrary files outside the intended extraction directory with the privileges of the extracting process via symbolic or hard links contained in the archive.
To remediate this issue, users should upgrade to version 0.17.0 or later.5hCVE-2026-1185174.3 MÉD—
——0The 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.7hCVE-2026-1142328.1 ALT—
——0A flaw was found in the UserManagedPermissionService of the Keycloak authorization server. The UMA Protection API, which manages user permissions, fails to validate the resourceType field when a user creates a new policy. This allows an attacker to create a policy that applies to resources owned by other users, potentially leading to unauthorized access or modification of sensitive data across the entire realm.5hCVE-2026-1122088.1 ALT—
——0A flaw was found in the Organization identity-first login flow of the Keycloak authentication server. Keycloak is an open-source identity and access management solution used to secure modern applications and services. The issue occurs when an attacker who belongs to multiple organizations initiates a login and then switches to a registration or password reset path. During this process, a flag indicating a successful login is not properly cleared when the user is removed from the session. An attacker can then provide a victim's username, and the system incorrectly treats the stale flag as proof of the victim's identity. This allows the attacker to bypass authentication and gain full access to the victim's account and connected applications.7hCVE-2026-1112536.6 MÉD—
——0A flaw was found in dnf5daemon-server. Session methods on dynamically exported D-Bus session objects do not verify that the caller is the unique sender that created the session. A local attacker can modify another user's live package-management session and replace its resolved transaction plan. If the legitimate session party later authorizes execution, the root daemon may apply the attacker-chosen package operations.7hCVE-2026-1109327.5 ALT—
——0A flaw was found in the theme resource endpoint of Keycloak, an open-source identity and access management solution. The issue occurs because the server fails to validate theme names before using them as keys for an on-disk gzip cache. An unauthenticated attacker can exploit this by sending many requests with unique, fabricated theme names, leading to persistent disk space exhaustion that survives server restarts and causes service disruptions.5hCVE-2026-1109317.5 ALT—
——0A flaw was found in the SAML 1.1 assertion parser of Keycloak, a tool used for identity and access management. The issue occurs when the parser handles a specific element called AuthorityBinding within an authentication statement. Due to an incomplete fix for a previous vulnerability, the parser can enter an infinite loop that consumes all available memory, causing the entire Keycloak service to crash and become unavailable to users.7hCVE-2026-1076027.5 ALT—
——0A flaw was found in the SAML 1.1 XML parsing logic of Keycloak, an open-source identity and access management solution. The issue occurs when the parser encounters unexpected elements within a security token's key information section, causing the processing thread to enter an infinite loop. An unauthenticated remote attacker could exploit this by sending specially crafted requests, leading to 100% CPU usage and a total denial of service that requires a server restart to resolve.7hCVE-2026-1071237.5 ALT—
——0A flaw was found in the SAML XML processing component of Keycloak, an open-source identity and access management solution. The issue occurs when the server parses XML documents containing excessively long dateTime or duration values. Because these strings are not properly validated before processing, the XML parser consumes excessive CPU time. An unauthenticated remote attacker can exploit this by sending specially crafted SAML requests to the server, leading to high CPU utilization and a denial of service for all users.7hCVE-2026-1048674.3 MÉD—
——0Immich is a high-performance self-hosted photo and video management solution. From 3.0.0 until 3.2.0, the GET /api/admin/users/{id}/calendar-heatmap route in server/src/controllers/user-admin.controller.ts uses the ordinary Permission.UserRead permission without an admin-route requirement, and UserAdminService.getCalendarHeatmap in server/src/services/user-admin.service.ts does not compare the authenticated user with the requested user identifier or perform a separate access check. Any authenticated non-admin user can therefore request another account's upload or capture-date activity heatmap for an arbitrary date range. AssetRepository.getCalendarHeatmap in server/src/repositories/asset.repository.ts filters by owner, date, and deletion state but not visibility, so the returned daily counts and total count include archived and locked assets. The route exposes counts and activity dates but does not return asset metadata or content, and it does not affect integrity or availability. This issue is fixed in version 3.2.0.5hCVE-2026-1031127.5 ALT—
——0A flaw was found in the authentication session handling of the keycloak-services component in Keycloak. The issue occurs because the system fails to verify if an authenticator is valid for the current authentication flow when switching between the standard authentication path and the reset credentials path. An authenticated attacker can exploit this by triggering a race condition during the credential reset process to associate their successful authentication state with another user's session. This can result in the attacker gaining full access to the victim's account and changing their password.7hCVE-2026-1027397.8 ALT—
——0ESPectre is a Wi-Fi motion sensing platform for ESP32. Prior to 3.0.0-rc1, Micro-ESPectre analysis tools, including tools/6_optimize_filter_params.py, tools/1_analyze_raw_data.py, and tools/10_train_ml_model.py, load NPZ datasets with NumPy pickle support enabled through micro-espectre/tools/csi_utils.py and a direct loader. A malicious dataset contributor or pull-request author can place an object-typed array in a .npz file under micro-espectre/data, causing pickle-controlled code to execute as the maintainer or developer when a documented analysis tool accesses the field. This issue is fixed in version 3.0.0-rc1.7hCVE-2026-852863.7 BAI—
——0HomeBox is a home inventory and organization system. Prior to 0.27.0, the POST /api/v1/users/login endpoint performs password-hash verification only when the submitted username matches an existing account, while unknown usernames return immediately. Although both paths return the same 401 unauthorized response, the processing-time discrepancy allows an unauthenticated remote attacker to distinguish registered email addresses; rate limiting only reduces the number of attempts per window. This issue is fixed in version 0.27.0.8hCVE-2026-738568.6 ALT—
——0ModSecurity is an open source, cross platform web application firewall (WAF) engine for Apache, IIS and Nginx. From 3.0.0 until 3.0.17, src/transaction.cc compares a response Content-Type value against SecResponseBodyMimeType entries case-sensitively. A mixed-case response media type that is semantically equivalent to a configured lowercase type can therefore be treated as unconfigured, causing ModSecurity to skip RESPONSE_BODY inspection. An unauthenticated requester can then receive response content that a phase 4 rule would otherwise block, including sensitive error or debug data. This issue is fixed in version 3.0.17.8hCVE-2026-70660——
——0stable-diffusion.cpp is a pure C/C++ library for running diffusion model (Stable Diffusion, Flux, Wan, Qwen Image, Z-Image, and more) inference. Prior to master-764-e22272e, the safetensors loader in src/model_io/safetensors_io.cpp trusts each tensor's data_offsets after checking only that nbytes equals end - begin. A crafted .safetensors file can declare a byte range beyond the file's data section, causing src/model_loader.cpp to consume an attacker-chosen length from an unchecked mapped-file offset. Loading the model can read adjacent process memory into tensor buffers or crash the process. This issue is fixed in master-764-e22272e.8hCVE-2026-706597.8 ALT—
——0stable-diffusion.cpp is a pure C/C++ library for running diffusion model (Stable Diffusion, Flux, Wan, Qwen Image, Z-Image, and more) inference. From master-742-1a13107 until master-763-4fcc6fe, IMatrixCollector::load_imatrix in src/runtime/imatrix.cpp reuses the values and counts vectors for entries with the same tensor name but resizes them only when they are empty. An untrusted .imatrix file can provide a small nval for the first entry and a larger nval for a second entry with the same name, causing the accumulation loop to write attacker-controlled floating-point and integer values beyond both heap buffers. Loading the file through the --imat-in conversion path can corrupt the heap or crash the process before a model is required. This issue is fixed in master-763-4fcc6fe.8h