LMDeploy is a toolkit for compressing, deploying, and serving large language models. Versions prior to 0.12.3 have a Server-Side Request Forgery (SSRF) vulnerability in LMDeploy's vision-language module. The `load_image()` function in `lmdeploy/vl/utils.py` fetches arbitrary URLs without validating internal/private IP addresses, allowing attackers to access cloud metadata services, internal networks, and sensitive resources. Version 0.12.3 patches the issue.

PackageKit is a a D-Bus abstraction layer that allows the user to manage packages in a secure way using a cross-distro, cross-architecture API. PackageKit between and including versions 1.0.2 and 1.3.4 is vulnerable to a time-of-check time-of-use (TOCTOU) race condition on transaction flags that allows unprivileged users to install packages as root and thus leads to a local privilege escalation. This is patched in version 1.3.5. A local unprivileged user can install arbitrary RPM packages as root, including executing RPM scriptlets, without authentication. The vulnerability is a TOCTOU race condition on `transaction->cached_transaction_flags` combined with a silent state-machine guard that discards illegal backward transitions while leaving corrupted flags in place. Three bugs exist in `src/pk-transaction.c`: 1. Unconditional flag overwrite (line 4036): `InstallFiles()` writes caller-supplied flags to `transaction->cached_transaction_flags` without checking whether the transaction has already been authorized/started. A second call blindly overwrites the flags even while the transaction is RUNNING. 2. Silent state-transition rejection (lines 873–882): `pk_transaction_set_state()` silently discards backward state transitions (e.g. `RUNNING` → `WAITING_FOR_AUTH`) but the flag overwrite at step 1 already happened. The transaction continues running with corrupted flags. 3. Late flag read at execution time (lines 2273–2277): The scheduler's idle callback reads cached_transaction_flags at dispatch time, not at authorization time. If flags were overwritten between authorization and execution, the backend sees the attacker's flags.

A flaw has been found in Tenda HG10 HG7_HG9_HG10re_300001138_en_xpon. This issue affects the function formRoute of the file /boaform/formRouting of the component Boa Service. This manipulation of the argument nextHop causes buffer overflow. It is possible to initiate the attack remotely. The exploit has been published and may be used.

Use after free in DevTools in Google Chrome prior to 147.0.7727.117 allowed a remote attacker who had compromised the renderer process to potentially perform a sandbox escape via a crafted HTML page. (Chromium security severity: High)

Reliance on untrusted inputs in a security decision in Microsoft Office allows an unauthorized attacker to bypass a security feature locally.

A vulnerability was identified in Tenda F456 1.0.0.5. The impacted element is the function fromP2pListFilter of the file /goform/P2pListFilter. The manipulation of the argument menufacturer/Go leads to buffer overflow. The attack is possible to be carried out remotely. The exploit is publicly available and might be used.

Deserialization of untrusted data in Microsoft Bing allows an unauthorized attacker to execute code over a network.

llama.cpp is an inference of several LLM models in C/C++. Prior to version b8492, the RPC backend's deserialize_tensor() skips all bounds validation when a tensor's buffer field is 0. An unauthenticated attacker can read and write arbitrary process memory via crafted GRAPH_COMPUTE messages. Combined with pointer leaks from ALLOC_BUFFER/BUFFER_GET_BASE, this gives full ASLR bypass and remote code execution. No authentication required, just TCP access to the RPC server port. This issue has been patched in version b8492.

ERB is a templating system for Ruby. Ruby 2.7.0 (before ERB 2.2.0 was published on rubygems.org) introduced an `@_init` instance variable guard in `ERB#result` and `ERB#run` to prevent code execution when an ERB object is reconstructed via `Marshal.load` (deserialization). However, three other public methods that also evaluate `@src` via `eval()` were not given the same guard: `ERB#def_method`, `ERB#def_module`, and `ERB#def_class`. An attacker who can trigger `Marshal.load` on untrusted data in a Ruby application that has `erb` loaded can use `ERB#def_module` (zero-arg, default parameters) as a code execution sink, bypassing the `@_init` protection entirely. ERB 4.0.3.1, 4.0.4.1, 6.0.1.1, and 6.0.4 patch the issue.

A vulnerability was identified in Linksys MR9600 2.0.6.206937. This affects the function BTRequestGetSmartConnectStatus of the file /etc/init.d/run_central2.sh of the component JNAP Action Handler. The manipulation of the argument pin leads to os command injection. The attack may be initiated remotely. The exploit is publicly available and might be used. The vendor was contacted early about this disclosure but did not respond in any way.