NEWSkillSec — elevating AI Skills security from malware detection to capability auditingSkillSecLearn more →
← Back to Intel Center
SECURITY INTEL · DAILY
Link copiedRSS

2026-07-16 Daily Security Intelligence

6 itemsTop severity 10.0 (Critical)CSSA 1 · CVE 2 · Poisoning 3

CSSA Exclusive Early Warning1

CSSA Exclusive Early Warning10.0 Critical

QwenPaw Console Malicious Plugin Upload Leading to Remote Code Execution

An attacker can exploit the plugin upload endpoint, which lacks strict validation, by submitting a crafted archive file containing malicious code. Due to the absence of a file type allowlist and sandbox isolation on the server side, the uploaded content is directly deserialized or dynamically executed, thereby triggering underlying command injection. This vulnerability is network-reachable and requires no prior authentication. Successful exploitation results in complete control over the host environment, leading to the compromise of core assets and disruption of business continuity.

Component
QwenPaw is an open-source framework for AI agent development and management. Its core architecture adopts a front-end/back-end separation design, wher…
Type
Unrestricted Upload of Dangerous File Types (CWE-434)
Repo
github.com/agentscope-ai/QwenPaw · Stars 22847
Remediation
  • It is recommended to enforce strict file extension allowlist validation on the upload endpoint and disable parsing of dangerous file types. Combined with storage isolation strategies that place files outside the web-accessible directory, a sandbox mechanism should be introduced to prevent malicious code from being loaded.

CVE Intelligence2

CVE-2026-61740CVSS 9.8 Critical2026-07-15

LightRAG API Key Bypass Vulnerability Leading to Unauthorized Access

In LightRAG prior to version 1.5.4, when deployed with LIGHTRAG_API_KEY set but AUTH_ACCOUNTS not set, the X-API-Key protection mechanism can be bypassed. Because lightrag/api/auth.py falls back to a hardcoded DEFAULT_TOKEN_SECRET, the /auth-status and /login endpoints can generate guest JWTs, and combined_dependency in lightrag/api/utils_api.py accepts a valid guest token before checking the API key. A remote unauthenticated attacker can invoke endpoints protected by combined_auth, including document read, upload, delete, graph structure modification, and query endpoints. This vulnerability has been fixed in version 1.5.4.

Component
LightRAG provides simple and fast Retrieval-Augmented Generation (RAG) capabilities.
Risks
  • Unauthorized Data Access: An attacker can read, upload, or delete sensitive documents
  • System Integrity Compromise: An attacker can modify graph structures and execute arbitrary queries
  • Remote Unauthenticated Attack: An attacker can remotely trigger the vulnerability without any credentials, completely bypassing the authentication mechanism
Source
NVD · CVE-2026-61740
Remediation
  • Immediately upgrade LightRAG to version 1.5.4 or laterG to version 1.5.4 or later
  • Ensure the AUTH_ACCOUNTS environment variable is properly configured at deployment time
  • Monitor API access logs to detect anomalous unauthenticated request behavior
CVE-2026-62378CVSS 9.0 Critical2026-07-16

RustFS Console PDF Preview Stored XSS Vulnerability Leading to Administrator Credential Leakage

In RustFS Console versions 0.1.7 through 0.1.10, the components/object/preview-modal.tsx and components/object/pdf-viewer.tsx components of the web administration console contain a Stored Cross-Site Scripting (Stored XSS, CWE-79) vulnerability due to insufficient extension-based PDF preview path validation. An attacker can upload a file containing malicious HTML content disguised with a .pdf extension; when an administrator previews the file in the console, the malicious script is executed. This vulnerability is a regression of CVE-2026-27822 and directly results in the exposure of sensitive administrator credentials including AccessKeyId, SecretAccessKey, and SessionToken. The vulnerability affects all RustFS Console users who have not updated to the patched version (fixed in 0.1.10). An attacker only needs to induce an administrator to access or preview the malicious file to trigger the vulnerability and steal high-privilege credentials.

Component
RustFS Console is the web administration console for the RustFS distributed file system, providing visual management capabilities including file management, preview, and system configuration.
Risks
  • Complete administrator credential leakage: An attacker can steal the administrator's AccessKeyId, SecretAccessKey, and SessionToken, thereby fully taking over file system administration privileges
  • Persistent stored attack: The malicious script is stored server-side, meaning any user who accesses the preview page—especially administrators—will be affected without requiring re-injection each time
  • Lateral movement and data tampering: Using the stolen credentials, an attacker can further access the backend storage to read, modify, or delete sensitive data, and even plant backdoors
Source
NVD · CVE-2026-62378
Remediation
  • Immediately upgrade RustFS Console to version 0.1.10 or later to apply the official patchon 0.1.10 or later to apply the official patch
  • Enforce strict content-type validation on uploaded files, relying not only on file extensions but also verifying file magic numbers
  • Enable sandbox isolation in the preview feature to prevent execution of any script code

Package Poisoning3

Package Poisoningnpm2026-07-16

selparsecss-selector@1.1.0 flagged as malicious

This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.

MD5
7b26c27085e26ae23da8d6bf7429d00b
Package Poisoningpypi2026-07-16

pylogora@0.7.8 flagged as malicious

This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.

MD5
c00b8275ed32a1a3de7b753688d74e78
Package PoisoningPyPI2026-07-16

qwen-asr-pvt@0.0.6 flagged as malicious

This version was found communicating with a malicious domain and executing malicious commands. Audit your dependencies and pin safe versions immediately.

MD5
50ec8bbc99da8482a31b1bd86b739a23
← Back to Intel CenterRSSMonitor these risks automatically with CleanSource SCA →