Nuvoton NPCT7xx TPM firmware is vulnerable to physical side-channel attacks enabling extraction of Elliptic Curve Diffie-Hellman (ECDH) keys through electromagnetic or timing analysis when an attacker has direct physical access to the TPM device. The vulnerability affects confidentiality of cryptographic material but does not enable code execution or availability attacks. No active exploitation has been publicly reported, and the attack requires specialized knowledge and physical proximity to the target system.
SQL injection in PostgreSQL logical replication ALTER SUBSCRIPTION REFRESH PUBLICATION allows authenticated local or network users with table creation privileges to execute arbitrary SQL queries with the publication subscriber's credentials. The attack is deferred until the next REFRESH PUBLICATION command is executed, requiring user interaction or scheduled maintenance. PostgreSQL 16.x, 17.x, and 18.x versions prior to 16.14, 17.10, and 18.4 respectively are vulnerable; earlier versions are unaffected. No public exploit code or active exploitation has been identified.
Improper ownership validation in MCP Registry prior to version 1.7.9 allows authenticated publishers to bind io.github.<user>/* namespaces to OCI images they do not control when the upstream OCI registry returns HTTP 429 rate-limit responses. The vulnerability bypasses the label-match ownership proof check, enabling namespace hijacking for users who publish through OCI-based MCP servers. Patch available in version 1.7.9.
{id}/pin endpoint, which incorrectly checks for read permission instead of write permission. This privilege escalation enables read-only users to perform a write operation (toggling is_pinned state) that should be restricted to users with explicit write access. The vulnerability is limited to the pin operation and does not permit modification of note content, title, or access grants. Publicly available proof-of-concept demonstrates the bypass across all shared notes with read access.
Server-Side Request Forgery (SSRF) in GitLab Enterprise Edition's virtual registry upstream feature allows an authenticated low-privileged user who controls a virtual registry upstream to direct the GitLab server to issue HTTP requests against internal hosts that would otherwise be inaccessible from the network perimeter. Affected versions span GitLab EE 18.8 through 18.11 across three patch trains. No public exploit exists and no active exploitation has been identified at time of analysis; EPSS probability stands at 0.01%, consistent with the narrow preconditions required.
Out-of-bounds memory write in Google Chrome's Skia graphics library (versions prior to 148.0.7778.168) enables attackers with an already-compromised renderer process to escalate privileges via malicious print files. This represents a sandbox escape vulnerability requiring high attack complexity and user interaction (printing or opening a crafted print file). While CVSS rates this 3.1 (Low), the real-world risk depends on the attacker already having achieved renderer compromise — making this a second-stage attack vector. No active exploitation confirmed (EPSS 0.05%, not in CISA KEV), but vendor patch available as of Chrome 148.0.7778.168.
Google Chrome on Linux versions prior to 148.0.7778.168 suffers from an out-of-bounds read vulnerability in GPU processing that enables cross-origin data leakage. An attacker who has already compromised the renderer process can craft malicious HTML to read sensitive memory beyond allocated bounds, exposing data from other origins that should be isolated by the browser's same-origin policy. EPSS exploitation probability is very low (0.03%, 10th percentile) and no active exploitation or public POC has been identified. Google rates this Medium severity and released patch 148.0.7778.168 to address the issue.
Cross-origin data leakage in Google Chrome on Android allows compromised renderer processes to access sensitive information from other sites via malicious HTML pages. Affects Chrome versions prior to 148.0.7778.168 on Android platforms. Attack requires high complexity (user interaction with crafted content) and prior renderer compromise, limiting practical exploitation. EPSS score of 0.03% (10th percentile) and SSVC assessment indicating no active exploitation align with the limited real-world risk profile despite the cross-origin information disclosure capability.
Cross-origin data leakage in Google Chrome on Windows via ANGLE graphics library allows renderer-compromised attackers to steal sensitive data from other origins through specially crafted web pages. Affects Chrome versions prior to 148.0.7778.168 on Windows platforms. EPSS probability of 0.03% (10th percentile) indicates low observed exploitation likelihood, with CISA SSVC confirming no active exploitation and non-automatable attack chain. Vendor patch released in Chrome 148.0.7778.168 stable channel update. Attack requires successful renderer process compromise as prerequisite, combined with user interaction, creating a chained exploitation scenario rather than standalone vulnerability.
Out-of-bounds memory write in Google Chrome's ANGLE graphics library (Windows only) enables attackers who have already compromised the renderer process to corrupt memory via specially crafted HTML pages. Chrome 148.0.7778.168 patches this type confusion vulnerability. Despite high Chromium severity rating, CVSS 3.1 scores only 3.1 due to prerequisite renderer compromise, high attack complexity, and required user interaction. EPSS 0.03% (10th percentile) and SSVC indicating no known exploitation suggest limited real-world risk at time of analysis.
Cross-origin data leakage in Google Chrome versions prior to 148.0.7778.168 occurs when an attacker who has already compromised the renderer process exploits an object corruption flaw in the Compositing component. The vulnerability requires user interaction with a malicious HTML page and high attack complexity to leak sensitive cross-origin data. Google has released a patch in Chrome 148.0.7778.168, and with EPSS at 0.03% (10th percentile) and no evidence of active exploitation (SSVC: none), this represents a medium-priority targeted threat rather than widespread exploitation risk.
CSS injection in SAP NetWeaver Application Server ABAP allows unauthenticated remote attackers to inject malicious Cascading Style Sheets into web pages served by the application, with exploitation requiring user interaction (clicking or accessing the affected page). The injected CSS executes in the victim's browser context, resulting in low-impact confidentiality loss; integrity and availability are not affected. CVSS 3.1 reflects the limited impact and high attack complexity required.
Use after free in GPU in Google Chrome prior to 148.0.7778.168 allowed a remote attacker who had compromised the renderer process to perform an out of bounds memory write via a crafted HTML page. (Chromium security severity: High)
Site Isolation bypass in Google Chrome on macOS allows attackers who have already compromised the renderer process to leak limited cross-origin data via malicious HTML in ReadingMode. Affects Chrome versions prior to 148.0.7778.168 on Mac only. EPSS score of 0.02% (6th percentile) indicates very low predicted exploitation probability. No active exploitation detected (not in CISA KEV), no public POC identified. CVSS 3.1 assigns Low severity despite High vendor severity rating due to requiring both renderer compromise and user interaction, with impact limited to confidentiality only.
Site Isolation bypass in Google Chrome prior to 148.0.7778.168 allows remote attackers who have already compromised the renderer process to access cross-site data via crafted HTML pages. The vulnerability affects Chrome's AI policy enforcement, enabling a second-stage attack after initial renderer compromise. Attack complexity is high, requiring both initial renderer compromise and user interaction. EPSS score of 0.02% indicates very low exploitation probability, and no active exploitation or public POC has been identified. Vendor patch is available in Chrome 148.0.7778.168.
dbt-mcp DefaultUsageTracker transmits unredacted MCP tool arguments-including raw SQL queries and credential-bearing --vars JSON-to dbt Labs telemetry by default without user opt-in. Affects dbt-mcp ≤1.17.0; tracking is enabled unless users explicitly set DBT_SEND_ANONYMOUS_USAGE_STATS=false or DO_NOT_TRACK=1 before installation, creating silent exfiltration of potentially sensitive database schema, credentials, and personally identifiable information. The vulnerability has been verified by proof-of-concept source code analysis and execution against dbt-mcp v1.15.1.
HCL AION uses basic authorization tokens for authentication, exposing credentials to interception or misuse if not transmitted over encrypted channels. The vulnerability affects authenticated local or adjacent network attackers with low privileges and user interaction, resulting in limited confidentiality impact. CVSS 3.0 reflects low severity, though the underlying authentication weakness may enable credential theft in environments with unencrypted internal traffic.
GitLab Enterprise Edition's instance-level approval rule editing prevention control can be bypassed by authenticated Maintainer-level users due to missing server-side authorization checks (CWE-862). Affected are all GitLab EE versions from 16.10 through before 18.9.7, 18.10 before 18.10.6, and 18.11 before 18.11.3. An attacker holding Maintainer permissions can modify or delete project approval rules even when an administrator has explicitly enabled the instance-level restriction designed to prevent such changes, undermining the integrity of merge request approval workflows. No public exploit identified at time of analysis, and EPSS is at the 1st percentile, indicating negligible opportunistic exploitation risk.
Merge request approval bypass in GitLab Enterprise Edition allows authenticated users to circumvent policy-enforced approval gates due to improper cleanup of orphaned approval policy records. Affected are all GitLab EE deployments running versions from 15.7 up to (but not including) 18.9.7, 18.10.6, and 18.11.3. No public exploit exists and no active exploitation has been confirmed; however, the integrity impact carries meaningful governance risk in regulated environments where merge approval policies serve as a compliance or change-management control.
HCL AION exposes sensitive information through URL parameters, allowing disclosure via browser history, server logs, and intermediary systems. The vulnerability requires adjacent network access, high interaction complexity, and authenticated user involvement, resulting in limited confidentiality impact with a CVSS score of 2.6. No active exploitation has been confirmed.
HCL AION stores sensitive information in browser auto-complete caches for certain input fields, potentially exposing credentials or other sensitive data to local attackers or through browser history under specific conditions. The vulnerability requires adjacent network access, high interaction complexity, and local user privilege, limiting real-world exploitation scope but posing risk in shared or compromised workstations.
libsixel versions prior to 1.8.7-r2 crash on memory allocation failure in sixel_decode_raw and sixel_decode functions due to incorrect NULL pointer validation, allowing local attackers to trigger denial of service under low-memory conditions. The vulnerability affects any application using libsixel's public decoding APIs when system memory pressure causes malloc to fail.
dbt MCP Server logs complete tool arguments including SQL queries and database credentials in plaintext to disk when file logging is enabled. Versions up to 1.17.0 write unredacted arguments from every tool invocation to dbt-mcp.log, with sensitive data such as raw SQL queries, credential-bearing vars payloads, and node selectors persisting indefinitely without automatic rotation. A local attacker with read access to the log file can extract credentials and SQL logic. Publicly available proof-of-concept demonstrates credential and PII extraction from log files.
Double-free vulnerability in PoDoFo 1.0.0 through 1.0.3 allows local attackers with user interaction to trigger heap corruption via failed digest operations in PDF signing routines, potentially causing denial of service. The vulnerability exists in compute_hash_to_sign() where EVP_DigestFinal failure causes buf to be freed twice, corrupting heap metadata. CVSS score is 2.5 (low severity) but exploitation requires local access and user interaction. Patched in version 1.0.4.
Blind Server-Side Request Forgery in Nextcloud News prior to 28.3.0-beta.1 allows low-privileged authenticated users to weaponize the feed URL submission feature - available via both the web interface and API - to coerce the Nextcloud server into making outbound HTTP requests to attacker-specified internal or private IP ranges, including localhost. Response content is never relayed back to the attacker, constraining this to a reconnaissance primitive rather than a data-exfiltration channel. No public exploit has been identified at time of analysis, EPSS sits at 0.04% (12th percentile), and CISA has not listed this in KEV, collectively indicating a low-probability real-world exploitation scenario.
HCL AION fails to configure security-related HTTP response headers, potentially reducing browser-based protections against cross-site scripting and other client-side attacks. The vulnerability requires adjacent network access, high interaction complexity, low privilege authentication, and user interaction to achieve limited confidentiality impact. CVSS score of 2.3 reflects minimal real-world risk under current attack conditions.