| CVE-2026-40976 | In certain circumstances, Spring Boot's default web security is ineffective allowing unauthorized access to all endpoints. For an application to be vulnerable, it must: be a servlet-based web application; have no Spring Security configuration of its own and rely on the default web security filter chain; depend on spring-boot-actuator-autoconfigure; not depend on spring-boot-health. If any of the above does not apply, the application is not vulnerable. Affected: Spring Boot 4.0.0–4.0.5; upgrade to 4.0.6 or later per vendor advisory. | critical | 2026-07-24 |
| CVE-2026-40975 | Values produced by ${random.value} are not suitable for use as secrets. ${random.uuid} is not affected. ${random.int} and ${random.long} should never be used for secrets as they are numeric values with a predictable range. Affected: Spring Boot 4.0.0–4.0.5 (fix 4.0.6), 3.5.0–3.5.13 (fix 3.5.14), 3.4.0–3.4.15 (fix 3.4.16), 3.3.0–3.3.18 (fix 3.3.19), 2.7.0–2.7.32 (fix 2.7.33); random value property source / weak PRNG for secrets. Versions that are no longer supported are also affected per vendor advisory. | high | 2026-07-24 |
| CVE-2026-40974 | Spring Boot's Cassandra auto-configuration does not perform hostname verification when establishing an SSL connection to Cassandra. Affected: Spring Boot 4.0.0–4.0.5 (fix 4.0.6), 3.5.0–3.5.13 (fix 3.5.14), 3.4.0–3.4.15 (fix 3.4.16), 3.3.0–3.3.18 (fix 3.3.19), 2.7.0–2.7.32 (fix 2.7.33); Cassandra SSL auto-configuration. Versions that are no longer supported are also affected per vendor advisory. | critical | 2026-07-24 |
| CVE-2026-40973 | A local attacker on the same host as the application may be able to take control of the directory used by `ApplicationTemp`. When `server.servlet.session.persistent` is set to `true` and the attack persists across application restarts, this may allow the attacker to read session information and hijack authenticated users or deploy a gadget chain and execute code as the application's user. Affected: Spring Boot 4.0.0–4.0.5 (fix 4.0.6), 3.5.0–3.5.13 (fix 3.5.14), 3.4.0–3.4.15 (fix 3.4.16), 3.3.0–3.3.18 (fix 3.3.19), 2.7.0–2.7.32 (fix 2.7.33); predictable temp directory / `ApplicationTemp` ownership verification. Versions that are no longer supported are also affected per vendor advisory. | high | 2026-07-24 |
| CVE-2026-40972 | An attacker on the same network as the remote application may be able to utilize a timing attack to discover information about the remote secret. In extreme circumstances this could result in the attacker determining the secret and uploading changed classes, thereby achieving remote code execution in the remote application. Affected: Spring Boot 4.0.0–4.0.5 (fix 4.0.6), 3.5.0–3.5.13 (fix 3.5.14), 3.4.0–3.4.15 (fix 3.4.16), 3.3.0–3.3.18 (fix 3.3.19), 2.7.0–2.7.32 (fix 2.7.33); DevTools remote secret comparison. Versions that are no longer supported are also affected per vendor advisory. | high | 2026-07-24 |
| CVE-2026-40971 | When configured to use an SSL bundle, Spring Boot's RabbitMQ auto-configuration does not perform hostname verification when connecting to the RabbitMQ broker. Affected: Spring Boot 4.0.0–4.0.5 (fix 4.0.6), 3.5.0–3.5.13 (fix 3.5.14) per vendor advisory. | critical | 2026-07-24 |
| CVE-2026-4073 | The pdfl.io plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'pdflio' shortcode in all versions up to, and including, 1.0.5. This is due to insufficient input sanitization and output escaping on the 'text' shortcode attribute. The output_shortcode() function directly concatenates the user-supplied $text variable into HTML output without applying esc_html() or any other escaping function. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. | medium | 2026-07-24 |
| CVE-2026-40714 | Dell PowerProtect Data Manager, versions prior to 20.2.0.0, contain(s) an Improper Input Validation vulnerability. A high privileged attacker with remote access could potentially exploit this vulnerability, leading to Elevation of privileges. | high | 2026-07-24 |
| CVE-2026-40712 | Dell PowerProtect Data Manager, versions prior to 20.2.0.0, contain(s) an Improper Input Validation vulnerability in the REST API. A high privileged attacker with remote access could potentially exploit this vulnerability, leading to Elevation of privileges. | critical | 2026-07-24 |
| CVE-2026-40691 | In Unbound 1.9.0 up to and including 1.25.1, when a DNSCrypt query is received over TCP, the routine that encrypts the reply in place fails to bound the reply length against the destination buffer size. The size clamp that protects the UDP path is not applied on the TCP path, so a reply larger than 65504 bytes is shifted forward by 48 bytes inside a buffer of capacity equal to 'msg-buffer-size', writing past the end of the heap allocation. A single malicious encrypted query crashes the resolver and lead to denial of service. This vulnerability needs Unbound to be compiled with DNSCrypt support ('--enable-dnscrypt') and the 'dnscrypt:' clause to be configured and enabled for the listening interfaces. | high | 2026-07-24 |
| CVE-2026-40688 | An out-of-bounds write vulnerability [CWE-787] vulnerability in Fortinet FortiWeb 8.0.0 through 8.0.3, FortiWeb 7.6.0 through 7.6.6, FortiWeb 7.4.0 through 7.4.11 may allow a remote privileged attacker to execute arbitrary code or command via crafted HTTP requests. | high | 2026-07-24 |
| CVE-2026-40683 | In OpenStack Keystone before 28.0.1, the LDAP identity backend does not convert the user enabled attribute to a boolean when the user_enabled_invert configuration option is False (the default). The _ldap_res_to_model method in the UserApi class only performed string-to-boolean conversion when user_enabled_invert was True. When False, the raw string value from LDAP (e.g., "FALSE") was used directly. Since non-empty strings are truthy in Python, users marked as disabled in LDAP were treated as enabled by Keystone, allowing them to authenticate and perform actions. All deployments using the LDAP identity backend without user_enabled_invert=True or user_enabled_emulation are affected. | high | 2026-07-24 |
| CVE-2026-4065 | The Smart Slider 3 plugin for WordPress is vulnerable to unauthorized access and modification of data due to missing capability checks on multiple wp_ajax_smart-slider3 controller actions in all versions up to, and including, 3.5.1.33. The display_admin_ajax() method does not call checkForCap() (which requires unfiltered_html capability), and several controller actions only validate the nonce (validateToken()) without calling validatePermission(). This makes it possible for authenticated attackers, with Contributor-level access and above, to enumerate slider metadata and create, modify, and delete image storage records by obtaining the nextend_nonce exposed on post editor pages. | medium | 2026-07-24 |
| CVE-2026-40622 | NLnet Labs Unbound 1.16.2 up to and including version 1.25.0 has a vulnerability of the 'ghost domain names' family of attacks that could extend the ghost domain window by up to one cached TTL configured value. Similar to other 'ghost domain names' attacks, an adversary needs to control a (ghost) zone and be able to query a vulnerable Unbound. A single client NS query can cause Unbound to overwrite the cached expired parent-side referral NS rrset with the child-side apex NS rrset and essentially extend the ghost domain window by up to one cached TTL configured value ('cache-max-ttl'). In configurations where 'harden-referral-path: yes' is used (non-default configuration), no client NS query is required since Unbound implicitly performs that query. Unbound 1.25.1 contains a patch with a fix that does not allow extension of TTLs for (parent) NS records regardless of their trust. | high | 2026-07-24 |
| CVE-2026-40564 | Files or Directories Accessible to External Parties, Server-Side Request Forgery (SSRF) vulnerability in Apache Flink Kubernetes Operator. The FlinkSessionJob jarURI is currently not validated so that it points to user-owned files or addresses. This lets a user with CR create permissions read files from the operator pod's filesystem and pull content from any backing store reachable through Flink's pluggable filesystem layer and access them through the submitted Flink job. Furthermore for fetching from http/https addresses there is currently no allowlist on the URI scheme, no host check, no IP-range restriction, and no protection against pointing the URI at internal or link-local addresses.This issue affects Apache Flink Kubernetes Operator: from 1.3.0 before 1.15.0. Users are recommended to upgrade to version 1.15.0, which fixes the issue. | medium | 2026-07-24 |
| CVE-2026-40560 | Starman versions before 0.4018 for Perl allows HTTP Request Smuggling via Improper Header Precedence. Starman incorrectly prioritizes "Content-Length" over "Transfer-Encoding: chunked" when both headers are present in an HTTP request. Per RFC 7230 3.3.3, Transfer-Encoding must take precedence. An attacker could exploit this to smuggle malicious HTTP requests via a front-end reverse proxy. | high | 2026-07-24 |
| CVE-2026-4051 | IBM Engineering Lifecycle Management 7.0.3, 7.1.0, and 7.2.0 could allow an attacker with administrative privileges to execute remote code due to exposed method that is not properly restricted. | high | 2026-07-24 |
| CVE-2026-40430 | Pronetiqs IntraVUE Versions 3.2.1a14 and prior have a plaintext storage of a password vulnerability that could expose cleartext credentials through the API. | high | 2026-07-24 |
| CVE-2026-40384 | An improper validation of the search parameter of the com_media files API endpoint leads to a path traversal vulnerability. | medium | 2026-07-24 |
| CVE-2026-40383 | An improper validation of user-supplied input leads to a local file inclusion vulnerability. | high | 2026-07-24 |
| CVE-2026-40331 | Masa CMS is an open source content management system. In versions 7.2.0 through 7.2.9, 7.3.0 through 7.3.14, 7.4.0 through 7.4.9, and 7.5.0 through 7.5.2, the unauthenticated JSON API accepts an altTable parameter that is stored via the setAltTable() method without validation or sanitization. This value is injected directly into a SQL FROM clause within feedGateway.cfc. An unauthenticated attacker can pass an arbitrary subquery into the altTable parameter to read sensitive data from any table in the database in a single HTTP request, including administrative credentials and password reset tokens. This issue has been fixed in versions 7.2.10, 7.3.15, 7.4.10, and 7.5.3. As a workaround, apply validation to the setAltTable function in core/mura/content/feed/feedBean.cfc to restrict input to simple alphanumeric table names, or disable the JSON API if it is not required. | critical | 2026-07-24 |
| CVE-2026-40330 | Masa CMS is an open source content management system. In versions 7.2.0 through 7.2.9, 7.3.0 through 7.3.14, 7.4.0 through 7.4.9, and 7.5.0 through 7.5.2, a SQL injection vulnerability exists in the beanFeed.cfc component within the getQuery function's handling of the sortDirection parameter. The parameter value is concatenated directly into SQL queries without sanitization or parameterization. An unauthenticated remote attacker can exploit this to extract sensitive information, modify or delete database records, or potentially achieve remote code execution on the underlying database server. This issue has been fixed in versions 7.2.10, 7.3.15, 7.4.10, and 7.5.3. As a workaround, use a WAF to block or restrict access to the beanFeed.cfc component, or deploy rules to detect SQL injection patterns targeting the sortDirection parameter. | critical | 2026-07-24 |
| CVE-2026-40329 | Masa CMS is an open source content management system. In versions 7.5.2 and earlier, a SQL injection vulnerability exists in the beanFeed.cfc component within the getQuery function's processing of the sortBy parameter. The application fails to properly sanitize or parameterize this input before incorporating it into dynamic SQL statements. An unauthenticated remote attacker can execute arbitrary SQL commands against the database, potentially gaining access to sensitive data, modifying or deleting records, or escalating privileges to administrative control. This issue has been fixed in versions 7.2.10, 7.3.15, 7.4.10, and 7.5.3. As a workaround, configure WAF rules to block malicious SQL patterns in the sortBy parameter sent to beanFeed.cfc. | critical | 2026-07-24 |
| CVE-2026-40291 | Chamilo LMS is an open-source learning management system. In versions prior to 2.0.0-RC.3, an insecure direct object modification vulnerability in the PUT /api/users/{id} endpoint allows any authenticated user with ROLE_STUDENT to escalate their privileges to ROLE_ADMIN by modifying the roles field on their own user record. The API Platform security expression is_granted('EDIT', object) only verifies record ownership, and the roles field is included in the writable serialization group, enabling any user to set arbitrary roles such as ROLE_ADMIN. Successful exploitation grants full administrative control of the platform, including access to all courses, user data, grades, and administrative settings. This issue has been fixed in version 2.0.0-RC.3. | high | 2026-07-24 |
| CVE-2026-40280 | Gotenberg is an API-based document conversion tool. In versions 8.30.1 and earlier, the default private-IP deny-lists for the --webhook-deny-list and --api-download-from-deny-list flags use a case-sensitive regular expression (^https?://) to match URL schemes. Because Go's net/url.Parse() normalizes the scheme to lowercase before establishing the outbound TCP connection, an attacker can bypass the deny-list by simply capitalizing part of the URL scheme (e.g., HTTP://, HTTPS://, or Http://). This allows unauthenticated requests to reach internal network services, including private IP ranges, loopback addresses, and cloud instance metadata endpoints such as HTTP://169.254.169.254/latest/meta-data/. This bypasses the same security control that was patched in CVE-2026-27018. This issue has been fixed in version 8.31.0. | high | 2026-07-24 |
| CVE-2026-4025 | The PrivateContent Free plugin for WordPress is vulnerable to Stored Cross-Site Scripting via the 'align' shortcode attribute in the [pc-login-form] shortcode in all versions up to, and including, 1.2.0. This is due to insufficient input sanitization and output escaping on the 'align' attribute. Specifically, the attribute value flows from the shortcode through pc_login_form() to pc_static::form_align(), where it is directly concatenated into an HTML class attribute without esc_attr() or any escaping. This makes it possible for authenticated attackers, with Contributor-level access and above, to inject arbitrary web scripts in pages that will execute whenever a user accesses an injected page. | medium | 2026-07-24 |
| CVE-2026-40110 | Jupyter Server is the backend for Jupyter web applications. In versions 2.17.0 and earlier, the Origin header validation uses Python's re.match() to check incoming origins against the allow_origin_pat configuration value. Because re.match() only anchors at the start of the string and does not require a full match, a pattern intended to match only a trusted domain (e.g., trusted.example.com) will also match any origin that begins with that domain followed by additional characters (e.g., trusted.example.com.evil.com). An attacker who controls such a domain can bypass the CORS origin restriction and make cross-origin requests to the Jupyter Server API from an untrusted site. This issue has been fixed in version 2.18.0. | high | 2026-07-24 |
| CVE-2026-40075 | OpenMRS Core is an open source electronic medical record system platform. In versions 2.7.8 and earlier and versions 2.8.0 through 2.8.5, the `/openmrs/moduleResources/{moduleid}` endpoint is vulnerable to a path traversal attack. The ModuleResourcesServlet constructs a filesystem path from user-controlled input without performing path boundary validation — the getFile() method concatenates the user-supplied path into an absolute filesystem path without calling normalize() or checking that the result stays within the allowed module resources directory. Because this endpoint serves static resources required for rendering the login page, it is not protected by authentication filters, allowing unauthenticated exploitation. An attacker can traverse directories and read arbitrary files from the server filesystem, including /etc/passwd and application configuration files containing database credentials. Successful exploitation requires the target deployment to run on Apache Tomcat versions prior to 8.5.31, where the ..; path parameter bypass is not mitigated by the container. Deployments on Tomcat 8.5.31 or later and Tomcat 9.0.10 or later are protected at the container level, though the underlying code defect remains. This issue has been fixed in versions after 2.7.8 (within the 2.7.x branch) and in version 2.8.6 and later. | high | 2026-07-24 |
| CVE-2026-40068 | In versions 2.1.63 through 2.1.83 of Claude Code, the folder trust determination logic used the git worktree commondir file without validating its contents. An attacker could craft a malicious repository with a commondir file pointing to a path the victim had previously trusted, causing Claude Code to bypass its trust confirmation dialog and immediately execute hooks defined in `.claude/settings.json`. Exploitation requires the victim to clone the malicious repository and run Claude Code within it, and the attacker must know or guess a path the victim had already trusted. This issue has been fixed in version 2.1.84. | high | 2026-07-24 |
| CVE-2026-40037 | OpenClaw before 2026.3.31 (patched in 2026.4.8) contains a request body replay vulnerability in fetchWithSsrFGuard that allows unsafe request bodies to be resent across cross-origin redirects. Attackers can exploit this by triggering redirects to exfiltrate sensitive request data or headers to unintended origins. | high | 2026-07-24 |
| CVE-2026-40036 | Unfurl before 2026.04 contains an unbounded zlib decompression vulnerability in parse_compressed.py that allows remote attackers to cause denial of service. Attackers can submit highly compressed payloads via URL parameters to the /json/visjs endpoint that expand to gigabytes, exhausting server memory and crashing the service. | high | 2026-07-24 |
| CVE-2026-40035 | Unfurl through 2025.08 contains an improper input validation vulnerability in config parsing that enables Flask debug mode by default. The debug configuration value is read as a string and passed directly to app.run(), causing any non-empty string to evaluate truthy, allowing attackers to access the Werkzeug debugger and disclose sensitive information or achieve remote code execution. | critical | 2026-07-24 |
| CVE-2026-40034 | gix-submodule before 0.29.0 (gitoxide before 0.5.21, gix before 0.84.0) incorrectly validates the update field in .gitmodules, allowing attackers to bypass the CommandForbiddenInModulesConfiguration guard when a submodule has been initialized with only partial configuration in .git/config. An attacker can inject arbitrary shell commands via the update field in .gitmodules that will be executed when Submodule::update() is called on a previously-initialized submodule, enabling remote code execution. | high | 2026-07-24 |
| CVE-2026-40033 | FreeRDP before 3.26.0 contains a heap-buffer-overflow vulnerability in gdi_CacheToSurface that allows remote attackers to write out-of-bounds heap memory. The vulnerability occurs because rectangle validation clamps coordinates to UINT16_MAX but performs copy operations using unclamped cache entry dimensions, enabling malicious RDP servers to trigger large out-of-bounds writes and potentially achieve remote code execution or client crash. | high | 2026-07-24 |
| CVE-2026-40032 | UAC (Unix-like Artifacts Collector) before 3.3.0-rc1 contains a command injection vulnerability in the placeholder substitution and command execution pipeline where the _run_command() function passes constructed command strings directly to eval without proper sanitization. Attackers can inject shell metacharacters or command substitutions through attacker-controlled inputs including %line% values from foreach iterators and %user% / %user_home% values derived from system files to achieve arbitrary command execution with the privileges of the UAC process. | high | 2026-07-24 |
| CVE-2026-40031 | MemProcFS before 5.17 contains multiple unsafe library-loading patterns that enable DLL and shared-library hijacking across six attack surfaces, including bare-name LoadLibraryU and dlopen calls without path qualification for vmmpyc, libMSCompression, and plugin DLLs. An attacker who places a malicious DLL or shared library in the working directory or manipulates LD_LIBRARY_PATH can achieve arbitrary code execution when MemProcFS loads. | high | 2026-07-24 |
| CVE-2026-40030 | parseusbs before 1.9 contains an OS command injection vulnerability where the volume listing path argument (-v flag) is passed unsanitized into an os.popen() shell command with ls, allowing arbitrary command injection via crafted volume path arguments containing shell metacharacters. An attacker can provide a crafted volume path via the -v flag that injects arbitrary commands during volume content enumeration. | high | 2026-07-24 |
| CVE-2026-4003 | The Users manager – PN plugin for WordPress is vulnerable to Privilege Escalation via Arbitrary User Meta Update in all versions up to and including 1.1.15. This is due to a flawed authorization logic check in the userspn_ajax_nopriv_server() function within the 'userspn_form_save' case. The conditional only blocks unauthenticated users when the user_id is empty, but when a non-empty user_id is supplied, execution bypasses this check entirely and proceeds to update arbitrary user meta via update_user_meta() without any authentication or authorization verification. Additionally, the nonce required for this AJAX endpoint ('userspn-nonce') is exposed to all visitors via wp_localize_script on the public wp_enqueue_scripts hook, rendering the nonce check ineffective as a security control. This makes it possible for unauthenticated attackers to update arbitrary user metadata for any user account, including the userspn_secret_token field. | critical | 2026-07-24 |
| CVE-2026-40029 | parseusbs before 1.9 contains an OS command injection vulnerability in parseUSBs.py where LNK file paths are passed unsanitized into an os.popen() shell command, allowing arbitrary command execution via crafted .lnk filenames containing shell metacharacters. An attacker can craft a .lnk filename with embedded shell metacharacters that execute arbitrary commands on the forensic examiner's machine during USB artifact parsing. | high | 2026-07-24 |
| CVE-2026-40028 | Hayabusa versions prior to 3.8.0 contain a cross-site scripting (XSS) vulnerability in its HTML report output that allows an attacker to execute arbitrary JavaScript when a user scans JSON-exported logs containing malicious content in the Computer field. An attacker can inject JavaScript into the Computer field of JSON logs that executes in the forensic examiner's browser session when viewing the generated HTML report, leading to information disclosure or code execution. | medium | 2026-07-24 |
| CVE-2026-40027 | ALEAPP (Android Logs Events And Protobuf Parser) through 3.4.0 contains a path traversal vulnerability in the NQ_Vault.py artifact parser that uses attacker-controlled file_name_from values from a database directly as the output filename, allowing arbitrary file writes outside the report output directory. An attacker can embed a path traversal payload such as ../../../outside_written.bin in the database to write files to arbitrary locations, potentially achieving code execution by overwriting executable files or configuration. | high | 2026-07-24 |
| CVE-2026-40026 | The Sleuth Kit through 4.14.0 contains an out-of-bounds read vulnerability in the ISO9660 filesystem parser where the parse_susp() function trusts len_id, len_des, and len_src fields from the disk image to memcpy data into a stack buffer without verifying that the source data falls within the parsed SUSP block. An attacker can craft a malicious ISO image that causes reads past the end of the SUSP data buffer, and a zero-length SUSP entry can trigger an infinite parsing loop. | medium | 2026-07-24 |
| CVE-2026-39907 | Unisys WebPerfect Image Suite versions 3.0.3960.22810 and 3.0.3960.22604 expose an unauthenticated WCF SOAP endpoint on TCP port 1208 that accepts unsanitized file paths in the ReadLicense action's LFName parameter, allowing remote attackers to trigger SMB connections and leak NTLMv2 machine-account hashes. Attackers can submit crafted SOAP requests with UNC paths to force the server to initiate outbound SMB connections, exposing authentication credentials that may be relayed for privilege escalation or lateral movement within the network. | high | 2026-07-24 |
| CVE-2026-39906 | Unisys WebPerfect Image Suite versions 3.0.3960.22810 and 3.0.3960.22604 expose a deprecated .NET Remoting TCP channel that allows remote unauthenticated attackers to leak NTLMv2 machine-account hashes by supplying a Windows UNC path as a target file argument through object-unmarshalling techniques. Attackers can capture the leaked NTLMv2 hash and relay it to other hosts to achieve privilege escalation or lateral movement depending on network configuration and patch level. | high | 2026-07-24 |
| CVE-2026-39901 | monetr is a budgeting application focused on planning for recurring expenses. Prior to 1.12.3, a transaction integrity flaw allows an authenticated tenant user to soft-delete synced non-manual transactions through the transaction update endpoint, despite the application explicitly blocking deletion of those transactions via the normal DELETE path. This bypass undermines the intended protection for imported transaction records and allows protected transactions to be hidden from normal views. This vulnerability is fixed in 1.12.3. | medium | 2026-07-24 |
| CVE-2026-39892 | cryptography is a package designed to expose cryptographic primitives and recipes to Python developers. From 45.0.0 to before 46.0.7, if a non-contiguous buffer was passed to APIs which accepted Python buffers (e.g. Hash.update()), this could lead to buffer overflows. This vulnerability is fixed in 46.0.7. | medium | 2026-07-24 |
| CVE-2026-39891 | PraisonAI is a multi-agent teams system. Prior to 4.5.115, the create_agent_centric_tools() function returns tools (like acp_create_file) that process file content using template rendering. When user input from agent.start() is passed directly into these tools without escaping, template expressions in the input are executed rather than treated as literal text. This vulnerability is fixed in 4.5.115. | high | 2026-07-24 |
| CVE-2026-39890 | PraisonAI is a multi-agent teams system. Prior to 4.5.115, the AgentService.loadAgentFromFile method uses the js-yaml library to parse YAML files without disabling dangerous tags (such as !!js/function and !!js/undefined). This allows an attacker to craft a malicious YAML file that, when parsed, executes arbitrary JavaScript code. An attacker can exploit this vulnerability by uploading a malicious agent definition file via the API endpoint, leading to remote code execution (RCE) on the server. This vulnerability is fixed in 4.5.115. | critical | 2026-07-24 |
| CVE-2026-39889 | PraisonAI is a multi-agent teams system. Prior to 4.5.115, the A2U (Agent-to-User) event stream server in PraisonAI exposes all agent activity without authentication. The create_a2u_routes() function registers the following endpoints with NO authentication checks: /a2u/info, /a2u/subscribe, /a2u/events/{stream_name}, /a2u/events/sub/{id}, and /a2u/health. This vulnerability is fixed in 4.5.115. | high | 2026-07-24 |
| CVE-2026-39888 | PraisonAI is a multi-agent teams system. Prior to 1.5.115, execute_code() in praisonaiagents.tools.python_tools defaults to sandbox_mode="sandbox", which runs user code in a subprocess wrapped with a restricted __builtins__ dict and an AST-based blocklist. The AST blocklist embedded inside the subprocess wrapper (blocked_attrs of python_tools.py) contains only 11 attribute names — a strict subset of the 30+ names blocked in the direct-execution path. The four attributes that form a frame-traversal chain out of the sandbox are all absent from the subprocess list (__traceback__, tb_frame, f_back, and f_builtins). Chaining these attributes through a caught exception exposes the real Python builtins dict of the subprocess wrapper frame, from which exec can be retrieved and called under a non-blocked variable name — bypassing every remaining security layer. This vulnerability is fixed in 1.5.115. | critical | 2026-07-24 |
