| CVE |
Vendors |
Products |
Updated |
CVSS v3.1 |
| adm-zip before 0.5.18 is vulnerable to denial of service via a crafted ZIP file with a manipulated uncompressed size header field. In zipEntry.js line 103, Buffer.alloc(_centralHeader.size) allocates memory based on the declared uncompressed size from the ZIP central directory header without validating it against the actual compressed data size or imposing any upper bound. The size value is read directly from the binary header at entryHeader.js line 266 with no bounds check. An attacker can craft a ~120-byte ZIP file that declares ~4GB uncompressed size, causing a memory allocation amplification ratio of over 33 million to 1. The allocation occurs before CRC validation, so the malicious payload cannot be rejected early. All extraction and read methods are affected: readFile(), readAsText(), extractEntryTo(), extractAllTo(), extractAllToAsync(), test(), and entry.getData(). Any application accepting untrusted ZIP files via adm-zip is vulnerable to immediate process crash. |
| Lima launches Linux virtual machines, typically on macOS, for running containerd. Prior to 2.1.3, on an instance of Lima running with the qemu driver, an arbitrary user in the VM could access /run/lima-guestagent.sock when the guest agent is enabled, which could result in running arbitrary commands with root privileges in the VM because the guest agent socket provides tunneling for arbitrary addresses, including Unix socket addresses for privileged daemons like D-Bus. This issue is fixed in version 2.1.3. |
| PraisonAI (pip package praisonaiagents) before 1.6.78 contains an unsafe dynamic module loading vulnerability in AgentFlow._resolve_pydantic_class (src/praisonai-agents/praisonaiagents/workflows/workflows.py). When a workflow step uses a string output_pydantic reference, the framework locates and imports a sibling tools.py from the workflow file's directory via importlib exec_module without sandboxing, ignoring the PRAISONAI_ALLOW_*_TOOLS environment variables. An attacker who controls a workflow file and its sibling tools.py can execute arbitrary Python code with the workflow runner's privileges when the workflow is executed via WorkflowManager or after load_yaml. |
| Capgo before 12.128.2 contains an authentication bypass vulnerability in the password change endpoint that allows attackers to change user passwords without requiring current password confirmation. Attackers with temporary session access can exploit this flaw to permanently lock out legitimate users and achieve full account takeover. |
| FlaskBB through 2.2.0, fixed in commit acc88cf, contains an authorization bypass vulnerability that allows authenticated moderators to perform unauthorized actions on topics in forums they do not control by submitting crafted topic ID lists. Attackers can include a low-ID topic from a permitted forum as an anchor in a batch request, causing the permission check applied only to the first result to pass, and then execute lock, unlock, delete, or hide actions against topics in unmoderated forums. |
| The SureForms – Drag and Drop Form Builder for WordPress plugin for WordPress is vulnerable to Improper Input Validation in all versions up to, and including, 2.2.1. This is due to the plugin accepting the payment amount directly from user-controlled POST data in the 'create_payment_intent' and 'create_subscription_intent' functions without validating it against the form's configured price. This makes it possible for unauthenticated attackers to modify the payment amount to any arbitrary value when submitting a Stripe payment form, potentially purchasing products or services at significantly reduced prices. |
| Langroid is a framework for building large-language-model-powered applications. Prior to version 0.64.0, Langroid's `ReadFileTool` and `WriteFileTool` appear to treat `curr_dir` as the intended working-directory boundary for file operations. However, the tools only change the process working directory to `curr_dir` and then operate on the user-supplied `file_path` without resolving and enforcing that the final path remains inside `curr_dir`. As a result, a tool caller can supply path traversal sequences such as `../secret.txt` to read files outside the configured current directory, or `../written_by_tool.txt` to write files outside that directory. This can impact applications that expose Langroid file tools to an LLM agent, user-controlled tool call, or delegated coding/documentation agent while relying on `curr_dir` to restrict file access to a project/workspace directory. Version 0.64.0 patches the issue. |
| Discourse is an open-source discussion platform. Prior to 2026.6.0, 2026.5.1, 2026.4.2, and 2026.1.5, a malicious second factor name on an attacker-controlled account was not escaped in the delete confirmation dialog, allowing stored cross-site scripting when an administrator impersonated that account. This issue is fixed in versions 2026.6.0, 2026.5.1, 2026.4.2, and 2026.1.5. |
| Discourse is an open-source discussion platform. Prior to 2026.6.0, 2026.5.1, 2026.4.2, and 2026.1.5, under certain non-default configurations, processing of PDF uploads could be exploited to obtain RCE on the server. This issue is patched in 2026.6.0, 2026.5.1, 2026.4.2, and 2026.1.5. |
| OpenClaw versions before 2026.6.6 contain a flaw in host exec environment filtering that can miss interpreter startup variables. When the affected feature is enabled and reachable, a lower-trust caller or configured input path can supply crafted environment variables to execute or persist actions beyond the caller's intended authorization. |
| OpenClaw versions before 2026.6.9 contain an authorization bypass vulnerability in the flock wrapper that allows lower-trust callers to execute or persist actions beyond their intended authorization. Attackers can leverage configured input paths to bypass durable exec approval binding and perform unauthorized operations when the affected feature is enabled. |
| A flaw was found in binutils. A heap-buffer-overflow vulnerability exists when processing a specially crafted XCOFF (Extended Common Object File Format) object file during linking. A local attacker could trick a user into processing this malicious file, which could lead to arbitrary code execution, allowing the attacker to run unauthorized commands, or cause a denial of service, making the system unavailable. |
| Laravel-Mediable before 7.0.0 contains a server-side request forgery vulnerability that allows remote attackers to issue arbitrary HTTP requests from the server by supplying unvalidated caller-controlled URLs to endpoints backed by MediaUploader::fromSource(). Attackers can craft URLs targeting RFC-1918 addresses, loopback interfaces, cloud metadata endpoints, or file:// URIs through RemoteUrlAdapter to reach internal infrastructure, retrieve sensitive files, and exfiltrate cloud credentials such as IAM tokens from instance metadata services. |
| Logto is the modern, open-source auth infrastructure for SaaS and AI apps. Prior to 1.41.0, Logto's self-hosted SAML application IdP built the signed SAML response and assertion by string-substituting user-controlled profile attributes such as name, email, and custom attribute-mapping values into element-text placeholders of a SAML XML template using samlify 2.10.0, which left those placeholders unescaped. An authenticated low-privilege user could place XML markup in a profile attribute so Logto signed a forged SAML attribute, such as an arbitrary role, allowing privilege escalation at relying Service Providers that authorize on SAML attributes. This issue is fixed in version 1.41.0. |
| OpenResty is a high performance web platform. From 1.29.2.1 to before 1.29.2.5, an out-of-bounds write vulnerability exists in the upstream PROXY protocol v2 implementation. When OpenResty is configured to send PROXY protocol version 2 headers to upstream servers, constructing the header in the stream proxy protocol v2 patch can write beyond the bounds of the allocated buffer, causing the worker process to crash and resulting in a denial of service. Only configurations that explicitly enable PROXY protocol v2 for upstream connections are impacted. This issue is fixed in version 1.29.2.5. |
| OpenReplay is a self-hosted session replay suite. In 1.27.0 and earlier, three dashboard and note mutation functions ran their SQL without the ownership predicate that their sibling read and edit functions use: notes.delete filtered only on note id and project id, while dashboards.update_widget and dashboards.remove_widget filtered only on dashboard id and widget id, allowing any authenticated member to delete another user's private session notes and remove or rewrite widgets on another user's private dashboards. |
| Software installed and run as a non-privileged user may cause OOB kernel memory reads or writes through GPU API calls.
When indexing pages larger than 4kB in the page freeing logic of the sparse memory implementation, incorrect buffer indexing leads to OOB access. |
| Software installed and run as a non-privileged user may conduct a sequence of improper GPU system calls causing use after free, which helps in facilitating unprivileged memory access from a shader code.
Triggering failure path in the MMU mapping logic by a malicious code could lead to incomplete cleanup of an internal driver state, allowing for future unauthorized access to the contents of the physical memory. |
| Kernel software installed and running inside a Host VM may post improper commands to the GPU Firmware to trigger a GPU register access which can lead to privilege escalation. |
| Software installed and run as a non-privileged user may conduct improper GPU system calls to cause an integer overflow and map two GPU virtual addresses to the same physical address. One of these virutal mappings can be freed along with the physical page, allowing for a read/write UAF via the second mapping
The second virtual mapping references a physical address that has been freed after the first virtual mapping has been freed. This allows the physical memory to be allocated (for example) by another process and read/written to. |
