| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| Flowise before 3.1.0 (affected versions 3.0.13 and earlier) uses weak hardcoded default JWT secrets ('auth_token', 'refresh_token') and default audience and issuer values ('AUDIENCE', 'ISSUER') in the enterprise passport authentication middleware (packages/server/src/enterprise/middleware/passport/index.ts). When the corresponding environment variables (JWT_AUTH_TOKEN_SECRET, JWT_REFRESH_TOKEN_SECRET, JWT_AUDIENCE, JWT_ISSUER) are not set, the application silently falls back to these publicly known defaults, allowing an attacker to forge valid JWTs and impersonate any user, including administrators, resulting in authentication bypass. |
| 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. |
| R-SOFT DMS is vulnerable to OS Command Injection in the Optical Character Recognition (OCR) module. Multiple command execution functions accept user-controllable file paths without proper sanitization before passing them to the system shell via SSH. In current infrastructure the URL encoding neutralizes the injection during the standard web upload flow. An authenticated attacker who is able to trigger the OCR functionality for the uploaded file can execute OS commands within the context of a root user.
This issue was fixed in version v3.19-2862 and v3.17-2580. |
| Certain devices in the WAGO System I/O Field series activate an internal diagnostic capability during the initial startup sequence. This functionality is not formally documented and becomes accessible without authentication for a brief period in the early boot phase. During this window, an unauthenticated remote attacker can gain access to the internal system processes, resulting in full system compromise. |
| The EVbee Service Android app uses TLS encrypted communication (HTTPS), but does not validate the certificate provided by the server. This allows an attacker on the network path between the app and EVbee server to intercept and manipulate the communication between the app and server. The traffic is weakly encrypted using RC4 with a hardcoded key, which allows an attacker to gain access to the communication. Part of this communication involves access codes to charging stations.
This issue affects EVbee Service: v1.4.101.00. |
| The network diagnosis endpoint on the web server at port 8090 is vulnerable to command injection. |
| The webserver running on port 8090 does not require authentication. This allows for sensitive information leakage such as configured passwords, or uploading files through different endpoints. |
| A POST request sent to a specific webserver endpoint can be used to write to arbitrary file locations. The endpoint accepts the filename parameter in the Content-Disposition header without verification.
This can be used to cause a denial of service by overwriting system files, or remote-code-execution by overwriting shell-scripts which execution can be triggered through other means. |
| The firmware update mechanism does not include cryptographic signature validation. This allows anyone with access to the firmware update capability to upload arbitrary files which can then lead to arbitrary code execution. |
| Various sensitive information such as passwords and charging card UIDs are written to log files. |
| The NPC start endpoint on the web server at port 8090 is vulnerable to command injection. |
| A vulnerability in Thales CERT "Suspicious" application =< 1.3.4 allows a remote and unauthenticated attacker to execute arbitrary code and arbitrarily overwrite writable application files—including Python modules, configuration files, cron inputs, and runtime artifacts—leading to a persistent denial of service, the potential compromise of application secrets or integrations, and root-level execution inside the Django application container.
This vulnerability has been names "Matryoshka Mail".
Thales PSIRT
acknowledges and thanks
Lucien Doustaly (aka wlayzz) for discovering and reporting this issue. |
| A Missing Authorization vulnerability in the repository creation functionality in Google Cloud BigQuery, Dataform and Colab Enterprise, in the versions between October 2025 and May 10th, 2026, on Google Cloud Platform, allows an authenticated attacker to escalate privileges and perform cross-tenant repository takeover.
This vulnerability was patched on 10 May 2026, and no customer action is needed. |
| Versions of Mura CMS prior to 10.0.712 contain a critical remote code execution (RCE) vulnerability. The flaw is located in the endpoint “/index.cfm/_api/json/v1/default”, where the “method” parameter in POST requests is not properly validated or sanitised before being processed by the ColdFusion engine. As a result, a remote attacker could exploit this vulnerability to inject and execute arbitrary CFML (ColdFusion Markup Language) expressions and instantiate malicious Java objects, thereby compromising the system’s security. |
| PraisonAI versions before 4.6.78 contain a code injection vulnerability in deploy/api.py where the agents_file parameter is directly interpolated into an f-string without sanitization. Attackers can inject arbitrary Python code that executes when the generated server code runs via subprocess.Popen(). |
| grav-plugin-database is the database plugin for Grav CMS. Prior to 1.2.0, the PDO::tableExists method interpolates its table argument directly into a raw SQL query string without sanitization, escaping, quoting, or whitelisting, allowing attacker-controlled table names passed by consuming plugin or developer code to execute arbitrary SQL against the configured database. This issue is fixed in version 1.2.0. |
| A vulnerability in the iCagenda extension for Joomla allows the upload of arbitrary files in the file attachment feature, ultimately resulting in PHP code upload and execution. |
| Vikunja before 2.2.1 contains an authorization flaw where the LinkSharing.ReadAll endpoint exposes share hashes to users with read access, enabling permission escalation to admin-level shares. The GetTaskAttachment endpoint performs permission checks against user-supplied task IDs but fetches attachments by sequential ID without verifying ownership, allowing attackers to download and delete all file attachments across all projects instance-wide. |
| Langroid is a framework for building large-language-model-powered applications. Prior to version 0.65.1, the `SQLChatAgent` SQL-injection mitigation, with default `allow_dangerous_operations=False`, combines a raw-text regex blocklist (`_DANGEROUS_SQL_PATTERNS`) with a `sqlglot` SELECT-only statement allowlist. The blocklist entries that target callable functions require the function name to be immediately followed by `\s*\(`. PostgreSQL accepts the same call with the name separated from `(` by a quoted identifier, an inline comment, or schema qualification. These forms evade the regex, still parse as `SELECT`, and execute the same PostgreSQL function. This restores the `pg_read_file` server-side file-read primitive that the prior CVE-2026-25879 / GHSA-pmch-g965-grmr fix was meant to block: the parent advisory fixed a missing `pg_read_file` blocklist entry, while this report shows that the added regex is bypassable. Version 0.65.1 fixes the issue. |
| Langroid is a framework for building large-language-model-powered applications. Prior to version 0.65.5, Neo4jChatAgent passes LLM-generated Cypher queries straight to the Neo4j driver with no validation, no statement-type allowlist, and no opt-out gate. The query text is influenceable by prompt injection (direct user input or indirect content the agent reads back via RAG), so an attacker who can influence the prompt can read or destroy all graph data and, when APOC or dbms.security procedures are enabled on the server, achieve OS-command and filesystem access. This is the same defect class and threat model as the SQLChatAgent prompt-to-SQL-to-RCE issue fixed in version 0.63.0 (CVE-2026-25879); that fix did not extend to the neo4j module. Version 0.65.5 contains a fix for the neo4j module. |
