Information Disclosure
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security.
How It Works
Information disclosure occurs when an application unintentionally exposes sensitive data that aids attackers in reconnaissance or directly compromises security. This happens through multiple channels: verbose error messages that display stack traces revealing internal paths and frameworks, improperly secured debug endpoints left active in production, and misconfigured servers that expose directory listings or version control artifacts like .git folders. APIs often leak excessive data in responses—returning full user objects when only a name is needed, or revealing system internals through metadata fields.
Attackers exploit these exposures systematically. They probe for common sensitive files (.env, config.php, backup archives), trigger error conditions to extract framework details, and analyze response timing or content differences to enumerate valid usernames or resources. Even subtle variations—like "invalid password" versus "user not found"—enable account enumeration. Exposed configuration files frequently contain database credentials, API keys, or internal service URLs that unlock further attack vectors.
The attack flow typically starts with passive reconnaissance: examining HTTP headers, JavaScript bundles, and public endpoints for version information and architecture clues. Active probing follows—testing predictable paths, manipulating parameters to trigger exceptions, and comparing responses across similar requests to identify information leakage patterns.
Impact
- Credential compromise: Exposed configuration files, hardcoded secrets in source code, or API keys enable direct authentication bypass
- Attack surface mapping: Stack traces, framework versions, and internal paths help attackers craft targeted exploits for known vulnerabilities
- Data breach: Direct exposure of user data, payment information, or proprietary business logic through oversharing APIs or accessible backups
- Privilege escalation pathway: Internal URLs, service discovery information, and architecture details facilitate lateral movement and SSRF attacks
- Compliance violations: GDPR, PCI-DSS, and HIPAA penalties for exposing regulated data through preventable disclosures
Real-World Examples
A major Git repository exposure affected thousands of websites when .git folders remained accessible on production servers, allowing attackers to reconstruct entire source code histories including deleted commits containing credentials. Tools like GitDumper automated mass exploitation of this misconfiguration.
Cloud storage misconfigurations have repeatedly exposed sensitive data when companies left S3 buckets or Azure Blob containers publicly readable. One incident exposed 150 million voter records because verbose API error messages revealed the storage URL structure, and no authentication was required.
Framework debug modes left enabled in production have caused numerous breaches. Django's DEBUG=True setting exposed complete stack traces with database queries and environment variables, while Laravel's debug pages revealed encryption keys through the APP_KEY variable in environment dumps.
Mitigation
- Generic error pages: Return uniform error messages to users; log detailed exceptions server-side only
- Disable debug modes: Enforce production configurations that suppress stack traces, verbose logging, and debug endpoints through deployment automation
- Access control audits: Restrict or remove development artifacts (
.git, backup files,phpinfo()) and internal endpoints before deployment - Response minimization: API responses should return only necessary fields; implement allowlists rather than blocklists for data exposure
- Security headers: Deploy
X-Content-Type-Options, remove server version banners, and disable directory indexing - Timing consistency: Ensure authentication and validation responses take uniform time regardless of input validity
Recent CVEs (12571)
Information disclosure in Foswiki versions up to 2.1.10 allows unauthenticated remote attackers to access sensitive data through the Changes/Viewfile/Oops component. Public exploit code exists for this vulnerability. Upgrading to version 2.1.11 or later resolves the issue.
Arbitrary host file exfiltration from Cloud Hypervisor VMM versions 34.0-50.0. CVSS 10.0. Patch available.
Feathersjs versions 5.0.39 and below store unencrypted HTTP headers in base64-encoded session cookies, allowing attackers with network access to decode and retrieve sensitive internal infrastructure details such as API keys, service tokens, and internal IP addresses. Authenticated users can exploit this vulnerability in deployments behind reverse proxies or API gateways to gain unauthorized access to sensitive information. A patch is available for affected installations.
When a DAG failed during parsing, Airflow’s error-reporting in the UI could include the full kwargs passed to the operators. If those kwargs contained sensitive values (such as secrets), they might be exposed in the UI tracebacks to authenticated users who had permission to view that DAG. [CVSS 6.5 MEDIUM]
RustDesk Client for Windows file transfer functionality allows local attackers with low-privileged code execution to read arbitrary files through symlink injection, potentially disclosing sensitive information with SYSTEM-level access. An attacker can exploit the Transfer File feature by uploading a specially crafted symbolic link to bypass access controls and access protected files on the target system. No patch is currently available for this vulnerability.
Certain Samsung MultiXpress Multifunction Printers may be vulnerable to information disclosure, potentially exposing address book entries and other device configuration information through specific APIs without proper authorization.
Global Facilities Management Software versions up to 20230721a contains a security vulnerability (CVSS 7.1).
Blank admin credentials allowed in device web management. Admin can set empty password, making device fully accessible.
HTTP Basic Authentication over unencrypted connections in the device's embedded web interface allows attackers on the same network to passively intercept and capture user credentials. This cleartext transmission of authentication data exposes administrative access to network-based eavesdropping attacks. The lack of HTTPS/TLS support creates a significant credential compromise risk for affected devices with no available patch.
Insertion of Sensitive Information Into Sent Data vulnerability in themeglow JobBoard Job listing job-board-light allows Retrieve Embedded Sensitive Data.This issue affects JobBoard Job listing: from n/a through <= 1.2.8. [CVSS 5.9 MEDIUM]
Connections versions up to 7.0 contains a vulnerability that allows attackers to obtain limited information when a single piece of internal metadata is returned (CVSS 3.5).
Seraphinite Solutions Seraphinite Accelerator seraphinite-accelerator is affected by missing authorization (CVSS 4.3).
uTLS versions 1.6.0 through 1.8.0 fail to properly mimic Chrome's cipher suite selection behavior when using GREASE ECH, randomly choosing ChaCha20 for encrypted client hello while consistently using AES for the outer handshake—a mismatch that does not occur in actual Chrome and creates detectable fingerprints. This inconsistency affects users relying on uTLS for fingerprinting resistance and could enable network observers to distinguish uTLS traffic from legitimate Chrome connections. A patch is available to correct the cipher suite selection logic.
Tanium addressed an insertion of sensitive information into log file vulnerability in TanOS. [CVSS 5.3 MEDIUM]
Tanium Interact logs sensitive information that authenticated users can access, potentially exposing confidential data through log file inspection. The vulnerability requires valid credentials and does not allow modification or service disruption, limiting its impact to information disclosure.
OpenClaw versions prior to 2026.2.15 allow authenticated administrators to write files outside the skill installation directory due to insufficient validation of the targetDir parameter during skill installation. An admin user could exploit this path traversal vulnerability to place malicious files in arbitrary locations on the system. A patch is available in version 2026.2.15 and later.
OpenClaw AI assistant versions prior to 2026.2.15 allow local authenticated users to access session transcripts across peer accounts in multi-user shared-agent deployments due to insufficient session targeting restrictions. Additionally, Telegram webhook mode may fail to properly validate per-account secrets, potentially allowing unauthorized webhook access. The vulnerability primarily impacts multi-user environments with untrusted peers, while single-user or trusted deployments face limited practical risk.
Windmill is an open-source developer platform for internal code: APIs, background jobs, workflows and UIs. Versions 1.634.6 and below allow non-admin users to obtain Slack OAuth client secrets, which should only be accessible to workspace administrators. [CVSS 2.7 LOW]
Libredesk is a self-hosted customer support desk application. Versions prior to 1.0.2-0.20260215211005-727213631ce6 fail to validate destination URLs for webhooks, allowing an attacker posing as an authenticated "Application Admin" to force the server to make HTTP requests to arbitrary internal destinations. This could compromise the underlying cloud infrastructure or internal corporate network...
Tanium addressed an insertion of sensitive information into log file vulnerability in Trends. [CVSS 6.5 MEDIUM]
OpenClaw versions prior to 2026.2.14 expose sensitive configuration secrets through the skills.status endpoint to clients with operator.read privileges, allowing authenticated attackers to retrieve raw credential values including Discord tokens. The vulnerability affects AI/ML deployments where read-scoped access is intended to be non-sensitive; affected users should upgrade to version 2026.2.14 or later and rotate any exposed Discord tokens.
FormaLMS 4.1.18 and earlier allows unauthenticated attackers to enumerate valid usernames through the password recovery endpoint by observing differential error messages. This user enumeration vulnerability could enable an attacker to build a list of active accounts for targeted attacks. No patch is currently available for this medium-severity issue.
httpsig-hyper versions prior to 0.0.23 fail to properly validate HTTP message digest headers due to improper use of Rust's matches! macro, allowing attackers to forge or modify message bodies without detection. This vulnerability affects applications using the library for HTTP signature verification, enabling attackers to bypass integrity checks on signed requests. A patch is available in version 0.0.23 and later.
Penpot before version 2.13.2 contains a path traversal vulnerability in the font creation endpoint that allows authenticated users with team edit permissions to read arbitrary files from the server filesystem. By supplying local file paths such as `/etc/passwd` as font data, attackers can retrieve sensitive files including system configuration, application secrets, and credentials. Public exploit code exists for this vulnerability, which could enable further server compromise depending on the Penpot process permissions.
Unauthenticated attackers can bypass access controls in Alfresco Content Services to retrieve sensitive files from protected directories such as WEB-INF through the /share/page/resource/ endpoint. This vulnerability exposes critical configuration data and credentials without requiring authentication or user interaction. No patch is currently available for this remotely exploitable issue affecting Alfresco deployments.
Missing authorization validation in Pterodactyl Wings prior to version 1.12.1 allows authenticated nodes to access and manipulate servers across different nodes without proper ownership verification. An attacker with a valid node secret token can retrieve sensitive installation scripts, alter server installation states, and modify transfer statuses for servers they should not have access to. The vulnerability requires network access and valid node credentials but carries high impact due to potential exposure of secrets and cross-node server manipulation.
strongMan's credential encryption uses a static initialization vector with AES-CTR mode, causing all database fields to be encrypted with identical key streams. An attacker with database access can leverage publicly stored certificates to derive the key stream and decrypt stored private keys and EAP secrets. No patch is currently available for this high-severity vulnerability affecting strongSwan management deployments.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
Dell PowerProtect Data Manager versions prior to 19.22 contain an incorrect privilege assignment flaw that allows remote attackers with low-level credentials to escalate their privileges on affected systems. The vulnerability requires network access and valid authentication but no user interaction, making it exploitable by insiders or attackers who have obtained legitimate credentials. No patch is currently available.
villatheme Sales Countdown Timer for WooCommerce and WordPress sctv-sales-countdown-timer is affected by php remote file inclusion (CVSS 7.5).
Dell Unisphere for PowerMax 10.2 contains a file path control vulnerability that allows authenticated remote attackers to disclose sensitive information. The vulnerability requires low-privileged credentials and network access but no user interaction, making it accessible to internal threats or compromised accounts. Currently no patch is available to remediate this issue.
Dell Unisphere for PowerMax versions 10.2 suffer from a path traversal vulnerability (CWE-73) that allows authenticated remote attackers to delete arbitrary files on affected systems. An attacker with low-level privileges can exploit this flaw without user interaction to achieve denial of service or system compromise. No patch is currently available for this high-severity issue (CVSS 8.1).
Dell Unisphere for PowerMax 10.2 contains a path traversal vulnerability that allows authenticated remote attackers to overwrite arbitrary files on the system. This HIGH severity flaw (CVSS 8.8) requires only low privileges and network access to exploit, potentially enabling complete system compromise. No patch is currently available for this vulnerability.
Metagauss EventPrime eventprime-event-calendar-management contains a security vulnerability (CVSS 5.3).
Ninja Tables plugin through version 5.2.5 inadvertently exposes sensitive data in outbound communications, allowing authenticated users to retrieve embedded sensitive information. This authenticated vulnerability affects all installations of the affected versions, though exploitation requires valid user credentials. No patch is currently available.
A weakness has been identified in xlnt-community xlnt up to 1.6.1. Impacted is the function xlnt::detail::decode_base64 of the file source/detail/cryptography/base64.cpp of the component Encrypted XLSX File Parser. [CVSS 3.3 LOW]
Cyreneadmin versions up to 1.3.0. contains a vulnerability that allows attackers to improper authorization (CVSS 4.3).
Local File Inclusion in Prodigy Commerce WordPress plugin <= 3.2.9.
The Virusdie - One-click website security plugin for WordPress is vulnerable to Sensitive Information Exposure in all versions up to, and including, 1.1.7. This is due to missing capability checks on the `vd_get_apikey` function which is hooked to `wp_ajax_virusdie_apikey`. This makes it possible for authenticated attackers, with Subscriber-level access and above, to retrieve the site's Virusdie API key, which could be used to access the site owner's Virusdie account and potentially compromis...
The Web Accessibility by accessiBe plugin for WordPress is vulnerable to Sensitive Information Exposure in all versions up to, and including, 2.11. This is due to the `accessibe_render_js_in_footer()` function logging the complete plugin options array to the browser console on public pages, without restricting output to privileged users or checking for debug mode. This makes it possible for unauthenticated attackers to view sensitive configuration data, including email addresses, accessiBe us...
Visual Integrated Command And Dispatch Platform versions up to 20260206. contains a security vulnerability (CVSS 6.5).
SD.NET RIM versions before 4.7.3c contain a SQL injection vulnerability that allows attackers to inject malicious SQL statements through POST parameters 'idtyp' and 'idgremium'. [CVSS 8.2 HIGH]
Visual Integrated Command And Dispatch Platform versions up to 20260206. contains a security vulnerability (CVSS 7.3).
Visual Integrated Command And Dispatch Platform versions up to 20260206. contains a security vulnerability (CVSS 5.3).
A flaw was found in QEMU in the uefi-vars virtual device. When the guest writes to register UEFI_VARS_REG_BUFFER_SIZE, the .write callback `uefi_vars_write` is invoked. [CVSS 3.3 LOW]
An insufficient entropy vulnerability was found in glibc. The getrandom and arc4random family of functions may return predictable randomness if these functions are called again after the fork, which happens concurrently with a call to any of these functions. [CVSS 4.8 MEDIUM]
OpenStack Nova compute nodes using the Flat image backend can have their host data destroyed when an authenticated user crafts a malicious QCOW header on a disk image and triggers a resize operation, causing qemu-img to execute without format restrictions. Affected versions include Nova before 30.2.2, 31.x before 31.2.1, and 32.x before 32.1.1, with no patch currently available. This vulnerability requires low privileges and user interaction but impacts the integrity and availability of the host system across trust boundaries.
Splunk Enterprise and Splunk Cloud Platform deployments expose SAML authentication configurations in plaintext logs accessible to users with Search Head Cluster administrative roles and _internal index access, allowing credential and authentication extension disclosure. Affected versions include Splunk Enterprise below 10.2.0, 10.0.2, 9.4.7, 9.3.8, and 9.2.11, as well as Splunk Cloud Platform versions below 10.2.2510.0, 10.1.2507.11, 10.0.2503.9, and 9.3.2411.120. No patch is currently available for this medium-severity vulnerability.
Splunk Enterprise versions before 10.2.0, 10.0.2, 9.4.7, 9.3.9, and 9.2.11 expose RSA access keys in plain text within the Authentication.conf file to users with access to the _internal index on Search Head Cluster deployments. A privileged user with appropriate role permissions could read these sensitive credentials, compromising authentication security. No patch is currently available for this medium-severity vulnerability.
Improper access control in Splunk Enterprise versions below 9.3.9, 9.4.8, and 10.0.2 allows low-privileged users without admin roles to access the Monitoring Console App endpoints, enabling unauthorized disclosure of sensitive information. The vulnerability affects only on-premises Splunk Enterprise deployments and does not impact Splunk Cloud Platform instances. No patch is currently available.
Splunk Search Head Cluster deployments below versions 10.2.0, 10.0.2, 9.4.7, 9.3.9, and 9.2.11 expose Duo Two-Factor Authentication secrets (integrationKey, secretKey, appSecretKey) in plain text to users with access to the _internal index and appropriate roles. An authenticated attacker with these privileges could retrieve sensitive credentials and compromise Duo authentication controls for the Splunk environment. No patch is currently available for this vulnerability.
Rejected reason: ** REJECT ** DO NOT USE THIS CANDIDATE NUMBER. No vendor patch available.
Rejected reason: ** REJECT ** DO NOT USE THIS CANDIDATE NUMBER. No vendor patch available.
Rejected reason: ** REJECT ** DO NOT USE THIS CANDIDATE NUMBER. No vendor patch available.
Linux kernel SMB client denial of service vulnerability caused by concurrent bitfield updates in the cached_fid structure that can corrupt flag states through read-modify-write races. A local attacker with standard privileges can trigger this race condition to cause availability disruptions by forcing inconsistent flag states in cached file handle management. No patch is currently available for this medium-severity issue.
The Linux kernel's virtio-crypto driver lacks proper synchronization when handling virtqueue notifications from multiple processes, causing data corruption and system hangs when processing cryptographic operations concurrently. Local attackers with user privileges can trigger denial of service by running parallel crypto workloads, as demonstrated through multi-process OpenSSL benchmarks that expose race conditions in the virtcrypto_done_task() handler. No patch is currently available for this medium-severity vulnerability affecting systems running virtio-crypto with builtin backends.
The Linux kernel ksmbd server leaks the active_num_conn counter when kthread_run() fails during TCP connection initialization, allowing local authenticated users to exhaust connection tracking resources and cause a denial of service. The vulnerability stems from improper cleanup that fails to decrement the counter when freeing the transport structure. No patch is currently available for this medium-severity issue.
The Exynos Virtual Display driver in the Linux kernel lacks proper synchronization when allocating and freeing memory structures, enabling use-after-free conditions through race conditions between concurrent operations. A local attacker with unprivileged access can exploit this vulnerability to cause memory corruption or achieve information disclosure by manipulating display connector operations. No patch is currently available for this high-severity vulnerability affecting Linux systems with Samsung Exynos graphics hardware.
The Linux kernel ksmbd subsystem contains a use-after-free vulnerability in multi-channel session handling due to missing synchronization on the ksmbd_chann_list xarray, allowing a local attacker with user privileges to cause memory corruption and potentially execute arbitrary code. The vulnerability affects the channel lookup and deletion operations between multiple concurrent sessions. A patch is available to add proper locking mechanisms to protect xarray access.
In the Linux kernel, the following vulnerability has been resolved: sched/mmcid: Don't assume CID is CPU owned on mode switch Shinichiro reported a KASAN UAF, which is actually an out of bounds access in the MMCID management code.
The Linux kernel erofs file system contains a use-after-free vulnerability in direct I/O file-backed mount operations that allows local attackers with user privileges to cause memory corruption and potentially achieve code execution or denial of service. The vulnerability occurs when accessing files through the directio option, where freed memory is subsequently accessed during I/O operations. A patch is not currently available, making this a critical concern for systems running affected Linux kernel versions.
Use-after-free vulnerability in Linux kernel XFS subsystem allows local attackers with unprivileged access to cause memory corruption and potential privilege escalation through improper pointer dereferencing in the btree block owner checking function. The flaw stems from attempting to access freed memory due to incorrect temporal ordering of operations when determining cursor aliases. This vulnerability affects all Linux systems using XFS and currently lacks a patch.
Memory allocation errors in the Linux kernel's OMAP crypto driver cause scatterlist objects to be undersized by 4x, enabling local authenticated attackers to trigger denial of service conditions through memory corruption. An attacker with local access and user-level privileges can exploit this miscalculation to crash the system or cause unpredictable kernel behavior. No patch is currently available for this vulnerability.
The Linux kernel's fsl-mc bus driver contains a use-after-free vulnerability in the driver_override_show() function that reads device configuration without proper locking, allowing a local privileged user to trigger memory corruption by concurrently modifying the same data. This vulnerability affects Linux systems running vulnerable kernel versions and could enable local denial of service or potential privilege escalation through heap memory manipulation. No patch is currently available for this issue.
In the Linux kernel, the following vulnerability has been resolved: crypto: iaa - Fix out-of-bounds index in find_empty_iaa_compression_mode The local variable 'i' is initialized with -EINVAL, but the for loop immediately overwrites it and -EINVAL is never returned.
In the Linux kernel, the following vulnerability has been resolved: hfs: ensure sb->s_fs_info is always cleaned up When hfs was converted to the new mount api a bug was introduced by changing the allocation pattern of sb->s_fs_info.
Insecure default telnet credentials in UTT HiPER 810 router firmware v1.5.0. Default credentials are publicly known, enabling unauthenticated access to the router management. PoC available.
A deadlock vulnerability in the Linux kernel's RISC-V tracing subsystem allows local users with tracing privileges to hang the system by enabling ftrace snapshots on __sbi_ecall functions, causing recursive IPI interrupts that trigger infinite snapshot loops. This issue is particularly easy to exploit on RISC-V systems lacking the SSTC extension, where timer events automatically invoke SBI ecalls. The vulnerability requires local access and is only exploitable if tracing is enabled, making it a denial of service vector for systems with active kernel tracing.
A use-after-free vulnerability in the Linux kernel's iSCSI target implementation allows local attackers with low privileges to cause memory corruption and potential denial of service by exploiting a race condition in the connection usage counting mechanism. The flaw occurs when a spinlock is released after calling complete(), allowing a waiting thread to free the connection structure before the current thread finishes its unlock operation. No patch is currently available for this vulnerability.
The Linux kernel's VMware hypercall implementation improperly handles register state during mouse events, allowing local attackers with user privileges to trigger a denial of service through a kernel panic via crafted input to the vmmouse driver. The vulnerability stems from incomplete register preservation when the QEMU VMware mouse emulation clears the upper 32 bits of CPU registers containing kernel pointers. No patch is currently available for this medium-severity issue affecting Linux systems running on VMware or QEMU with vmmouse support.
A denial of service vulnerability in the Linux kernel's btrfs filesystem allows local users with standard privileges to cause a system crash by triggering transaction aborts on read-only mounted filesystems. An attacker can exploit this by mounting a malformed btrfs filesystem with rescue options, causing the kernel to abort transactions with error handling failures during unmount. No patch is currently available for this medium-severity vulnerability.
AMD GPU drivers on Linux systems fail to prevent MMIO register access during SMU Mode 1 reset, allowing incomplete PCIe transactions that can trigger NMI panics or system hangs. A local attacker with driver interaction capabilities could exploit this to cause a denial of service by accessing registers while the device is offline. The vulnerability affects Linux kernel implementations with AMD PM functionality and currently lacks an available patch.
A data-race condition in the Linux kernel bonding driver's slave->last_rx field can be accessed without proper synchronization, potentially causing a denial of service on systems using bonded network interfaces. Local attackers with limited privileges can trigger the race condition to cause system instability or crashes. A patch is not currently available, and exploitation requires specific timing conditions.
Linux kernel swap handling can cause a kernel panic under heavy memory pressure when arch_prepare_to_swap fails due to read-only swap address space restrictions introduced in a prior commit. A local attacker with user privileges can trigger this denial of service condition during memory reclamation operations. No patch is currently available for this medium-severity vulnerability.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
In the Linux kernel, the following vulnerability has been resolved: wifi: mac80211: don't WARN for connections on invalid channels It's not clear (to me) how exactly syzbot managed to hit this, but it seems conceivable that e.g.
Rejected reason: This CVE ID has been rejected or withdrawn by its CVE Numbering Authority. No vendor patch available.
In the Linux kernel, the following vulnerability has been resolved: md: suspend array while updating raid_disks via sysfs In raid1_reshape(), freeze_array() is called before modifying the r1bio memory pool (conf->r1bio_pool) and conf->raid_disks, and unfreeze_array() is called after the update is completed.
NVIDIA NeMo Framework contains a vulnerability where an attacker could cause remote code execution by convincing a user to load a maliciously crafted file. A successful exploit of this vulnerability might lead to code execution, denial of service, information disclosure, and data tampering. [CVSS 7.8 HIGH]
NVIDIA NeMo Framework contains a vulnerability where an attacker could cause remote code execution. A successful exploit of this vulnerability might lead to code execution, denial of service, information disclosure, and data tampering. [CVSS 7.8 HIGH]
NVIDIA NeMo Framework contains a vulnerability where an attacker could cause remote code execution. A successful exploit of this vulnerability might lead to code execution, denial of service, information disclosure, and data tampering. [CVSS 7.8 HIGH]
NVIDIA NeMo Framework contains a vulnerability where an attacker could cause remote code execution. A successful exploit of this vulnerability might lead to code execution, denial of service, information disclosure, and data tampering. [CVSS 7.8 HIGH]
NVIDIA NeMo Framework for all platforms contains a vulnerability in a voice-preprocessing script, where malicious input created by an attacker could cause a code injection. [CVSS 7.8 HIGH]
NVIDIA NeMo Framework for all platforms contains a vulnerability in the ASR Evaluator utility, where a user could cause a command injection by supplying crafted input to a configuration parameter. [CVSS 7.8 HIGH]
NVIDIA NeMo Framework contains a vulnerability where malicious data could cause remote code execution. A successful exploit of this vulnerability might lead to code execution, escalation of privileges, information disclosure, and data tampering. [CVSS 8.0 HIGH]