Centreon Web, an open-source IT monitoring platform, contains an SQL injection flaw in its ACL management (Resources & Actions). An authenticated user with ACL management rights can read and modify the Centreon database. The attacker needs high privileges, so the main risk is a malicious or compromised administrator. The database stores monitoring configuration and credentials for monitored systems.
Centreon · Centreon Web
The issuer resolution logic in the AWS Private CA Connector for Kubernetes (aws-privateca-issuer) before v1.10.1 does not correctly enforce authorization. An authenticated user whose approval policy only permits a namespace-scoped issuer may obtain certificates, including subordinate CA certificates, signed by a cluster-scoped AWS Private CA issuer they are not authorized to use. The attack requires valid cluster credentials and low privileges.
Amazon Web Services · AWS Private CA Connector for Kubernetes
Botan is a C++ cryptography library. In versions 3.7.0 up to 3.13.0, AutoSeeded_RNG can mark itself as seeded without consulting the configured entropy source if an application clears the generator and then requests empty output while it is unseeded. On builds without a system RNG, later non-empty requests can then produce output from predictable state, which may weaken keys, nonces or tokens. The attack vector is local with low privileges, and exploitation is not known.
randombit · Botan
The npm package minimatch, which converts glob patterns into JavaScript RegExp objects, can be driven into excessive CPU and memory consumption during pattern compilation. A short, attacker-controlled glob pattern containing **/.. segments triggers repeated expansion at optimizationLevel 2 or higher. The impact is limited to availability, so a vulnerable application may become slow or crash.
isaacs · minimatch
The @sigstore/oci library of sigstore-js sends registry Authorization headers, Docker configuration headers or Basic credentials to an upload location supplied by the registry, without checking that it matches the registry's origin. A malicious or compromised registry can redirect a user pushing an artifact via attachArtifactToImage() to an attacker-controlled host and obtain the bearer token or credentials. The attack requires the victim to push an artifact and a malicious or compromised registry; the impact is limited to confidentiality.
Sigstore · sigstore-js (@sigstore/oci)
The format-fixing GitHub Actions workflow of the Monkeytype repository runs on pull_request_target with write permissions and executes code from a fork pull request. An external contributor whose pull request receives the format label can run arbitrary code, read the GITHUB_TOKEN and push repository content or modify pull requests and issues. The typing test application and user data are not affected.
monkeytypegame · Monkeytype
Kubeflow Pipelines API server versions before 2.18.0 decompress uploaded pipeline archives without sufficiently bounding the expanded size. A small, specially crafted ZIP or gzip-compressed tar archive can consume large amounts of memory and exhaust API server resources. Because decompression happens before namespace authorization, any caller that can reach the upload endpoint can trigger it without valid credentials.
Kubeflow · Kubeflow Pipelines
Research and translation in progress – available shortly.
Kubeflow Pipelines before 2.18.0 does not check whether the caller may access a run before returning its pod logs through the API server's run log endpoint. An attacker who can reach the API server and knows a run ID and a suitable pod or node identifier can read live or archived logs with the API server's permissions. Because the live log path also accepts the node ID as a pod name without verifying the run, a known run ID can be used to read logs of an unrelated pod in another namespace. Exposed logs may contain sensitive workload data or credentials.
Kubeflow · Pipelines
Kubeflow Pipelines before 2.18.0 does not restrict the file path supplied to the GET /artifacts/get?source=volume route to the selected volume. An attacker who can reach this route and knows the name of an existing mounted volume can read files outside that volume, as far as the artifact-serving process has read access. With authentication disabled, no credentials are required; in multi-user deployments the caller normally must be authorized for the requested namespace. The impact is limited to confidentiality.
Kubeflow · Pipelines
Rapid7 Insight Agent on Linux and macOS does not handle local filesystem paths correctly. An authenticated local user with low privileges can gain elevated privileges when the agent's inventory collection processes a crafted path. Exploitation requires an existing local account and is not network-reachable.
Rapid7 · Insight Agent
Tailscale Android build versionCode 297862600 exports a broadcast receiver without a permission or caller check. A locally installed app with no permissions can send a broadcast that changes the VPN exit-node setting, the allow-LAN-access setting, or the connect/disconnect state of the VPN. The exit node must already exist in the victim's netmap, and the connect action may require the system VPN consent on a fresh profile. The CVE record itself describes the issue only as insecure permissions, and the CVSS 9.1 score and CWE-732 are from NVD and could not be confirmed on NVD at the time of research.
Tailscale · Android app
The Nextcloud Android app (com.nextcloud.client) contains an exported activity without permission protection. A malicious app installed on the same phone, holding no permissions, can send a crafted login deep link that creates an account and sets it as current, so the victim app switches to an attacker-controlled server profile. The attacker must control the server named in the link, and no user interaction is needed. The NVD record and its network-based CVSS vector are unconfirmed, and the third-party advisory describes a local attack, so the severity rating is disputed.
Nextcloud · Android client
Shelter versionCode 445 (F-Droid version 1.9.1) exports an activity that accepts intents from any app without a permission. According to a third-party advisory, a local app with no permissions can register its own auth_key on a Shelter instance that was never enrolled and then use the app's binder service and change its preference state. Once a key is enrolled, the stored key is required, so the main risk is during first setup. The NVD score of 9.8 assumes network access, which the advisory does not confirm, so this should be treated as a risk from other apps on the same device.
PeterCxy · Shelter
AndBible, an offline Bible study app for Android, is affected by an issue in versionCode 1111 that allows arbitrary code execution. An attacker can exploit it by getting a user to upload a crafted module zip file. The CVE record does not state whether any user interaction beyond the upload is required, and no authentication is mentioned. The CVSS 9.8 score and the CWE classification in the CVE data could not be verified against NVD.
AndBible · AndBible: Bible Study
The io.file.importer.ImportActivity component of OpenTracks versionCode 6751 has insecure permissions. An attacker can inject crafted tracking data into the location database of the app via a crafted SEND message. The records were reported with CVSS 9.1 by the CVE data, but the scoring and attack vector are not confirmed by other sources, and the advisory text could not be read directly.
OpenTracks Team · OpenTracks
neutriNote CE for Android, versionCode 1387, has insecure permissions in its autosave component. According to the linked third-party advisory, an app without permissions could launch an exported activity and make the app write attacker-chosen content to a file URI the app can access. The CVE record states that unauthorized attackers can write arbitrary files, but the conditions and the real attack path are not confirmed. The advisory's CVSS attack vector (network) is disputed, and the sources do not show that the flaw is reachable remotely.
AppML · neutriNote CE
Librera Reader for Android (versionCode 7222) exports an activity that other apps can call without any permission. A calling app can supply a crafted file name that the reader uses to write and truncate files in shared storage outside its sandbox. No user interaction is needed, and the advisory states that writes beyond the Download folder additionally require that the victim app has been granted All files access.
foobnix · Librera Reader
The AIDL interface of the MAXS Module FileRead Android add-on (versionCode 27) has insecure permissions. An unauthorized attacker can read arbitrary files outside the app sandbox by supplying an absolute path. The CVSS 3.1 base score of 7.5 comes from NVD; VulDB rates the issue at 5.3 and limits the impact to confidentiality, and the required access path is unclear.
ProjectMAXS · MAXS Module FileRead
AndFHEM 2.4.1, an Android client for the FHEM home-automation server, exports two components without a permission check. Any app installed on the same device can send commands to the configured FHEM server, using the stored credentials of the user. The attack requires no user interaction, but the victim must have configured an FHEM server.
klassm · AndFHEM
RetroArch for Android (com.retroarch) version 1.22.2 exports an activity that accepts a file path from another app and loads that file as a native library. A locally installed app without any permissions can therefore make RetroArch copy and load attacker-supplied code, which runs with the uid and privileges of RetroArch. The official CVE record rates exploitation as possible through a proof of concept and automatable, but the NVD analysis is pending and the CVSS score of 9.8 is not yet confirmed by NVD.
libretro · RetroArch
An issue in Forkyz, an unofficial Android crossword app, versionCode 8500000 (Version 85), allows an attacker to change another user's storage settings by supplying a crafted URI string. The conditions for exploitation are not fully confirmed, and the vulnerability is unclear about whose storage is affected. The NVD score of 7.5 is not verified and conflicts with a VulDB score of 3.5 (Low).
Forkyz · Forkyz
Motion UI for Android (app.motionui.android), versionCode 101, exposes an activity to other apps on the same device without permission checks. A locally installed app can make it load an attacker-chosen URL in a WebView, which injects the victim's stored username and password into that page and submits them automatically. The attacker can receive the credentials and run script inside the WebView. The NVD description instead speaks of remote code execution via a crafted POST request; the primary advisory does not confirm this, so the NVD CVSS 9.8 rating should be treated with caution.
MotionCreative · Motion UI
GPSLogger version 137 for Android has insecure permissions that let unauthorized attackers manipulate the app's logging mechanism and the logged data. According to the third-party advisory, any app can start an exported activity without permissions and make GPSLogger download a profile from an attacker-chosen URL and activate it, which can redirect where logged data is uploaded. The victim must already have granted location access to the app. The CVSS vector assumes network access, while VulDB describes local access as required, so the real attack path needs verification.
mendhak · GPSLogger
Research and translation in progress – available shortly.
