Latest News

NVIDIA's incomplete security patch, combined with a Docker vulnerability, creates a serious threat for organizations using containerized environments. This article explains the risks and mitigation strategies.

Meta users in Europe will have their public posts swept up and ingested for AI training, the company announced this week.

Microsoft held off on releasing the privacy-unfriendly feature after a swell of pushback last year. Now it’s trying again, with a few improvements that skeptics say still aren't enough.

Prodaft is currently buying accounts from five Dark Web forums and offers to pay extra for administrator or moderator accounts. The idea is to infiltrate forums to boost its threat intelligence.

### Summary Unsanitized search param names cause XSS vulnerability. You are affected if you iterate over all entries of `event.url.searchParams` inside a server `load` function. Attackers can exploit it by crafting a malicious URL and getting a user to click a link with said URL. ### Details SvelteKit tracks which parameters in `event.url.searchParams` are read inside server `load` functions. If the application iterates over the these parameters, the `uses.search_params` array included in the boot script (embedded in the server-rendered HTML) will have any search param name included in unsanitized form. `packages/kit/src/runtime/server/utils.js:150` has the `stringify_uses(node)` function which prints these out. ### Reproduction In a `+page.server.js` or `+layout.server.js`: ```js /** @type {import('@sveltejs/kit').Load} */ export function load(event) { const values = {}; for (const key of event.url.searchParams.keys()) { values[key] = event.url.searchParams.get(key); ...

Cheap Android phones with preinstalled malware use fake apps like WhatsApp to hijack crypto transactions and steal wallet recovery phrases.

DevDojo Voyager 1.4.0 through 1.8.0, when Laravel 8 or later is used, allows authenticated administrators to execute arbitrary OS commands via a specific php artisan command.

### Impact Hook scripts in pleezer can be triggered by various events like track changes and playback state changes. In affected versions, these scripts were spawned without proper process cleanup, leaving zombie processes in the system's process table. Even during normal usage, every track change and playback event would leave behind zombie processes. This leads to inevitable resource exhaustion over time as the system's process table fills up, eventually preventing new processes from being created. The issue is exacerbated if events occur rapidly, whether through normal use (e.g., skipping through a playlist) or potential manipulation of the Deezer Connect protocol traffic. This vulnerability affects all users who have configured hook scripts using the `--hook` option. ### Patches This issue has been fixed in version 0.16.0. Users should upgrade to this version, which properly manages child processes using asynchronous process handling and cleanup. ### Workarounds Users who canno...

### Summary: A user with permission to create/modify EventSource and Sensor custom resources can gain privileged access to the host system and cluster, even without having direct administrative privileges. ### Details: The `EventSource` and `Sensor` CRs allow the corresponding orchestrated pod to be customized with `spec.template` and `spec.template.container` (with type `k8s.io/api/core/v1.Container`), thus, any specification under `container` such as `command`, `args`, `securityContext `, `volumeMount` can be specified, and applied to the EventSource or Sensor pod due to the code logic below. ```golang if args.EventSource.Spec.Template != nil && args.EventSource.Spec.Template.Container != nil { if err := mergo.Merge(&eventSourceContainer, args.EventSource.Spec.Template.Container, mergo.WithOverride); err != nil { return nil, err } } ``` With these, A user would be able to gain privileged access to the cluster host, if he/she specified the Even...

A threat actor posted about the zero-day exploit on the same day that Fortinet published a warning about known vulnerabilities under active exploitation.