CVE-2026-96512 is a local authorization bypass vulnerability in sudo on Red Hat Enterprise Linux 10. It occurs when sudoers rules use NOTBEFORE or NOTAFTER time-based restrictions with timestamps missing the trailing 'Z' timezone indicator. This causes sudo to rely on the TZ environment variable from the calling user, allowing an unprivileged user to manipulate the time window by setting TZ to an extreme offset and execute commands outside the intended time frame. Authentication is not bypassed; only the time-based authorization check is affected. This vulnerability requires a sudoers rule with the affected timestamp format and the attacker to be subject to that rule. Red Hat has released updates addressing this issue.
FastGPT is an open-source LLM platform for building AI applications on a knowledge base. Prior to 4.15.2, the safe Axios request interceptor in packages/service/common/api/axios.ts validates a hostname with isInternalAddress() before a later HTTP connection performs an independent DNS lookup, creating a DNS rebinding window, allowing an attacker-controlled hostname to resolve publicly during the check and to a loopback, private, link-local, or metadata address during connection. An authenticated attacker who can supply a URL to a safe-Axios-backed HTTP tool, workflow HTTP node, external file fetch, or other server-side integration can use this time-of-check/time-of-use gap to reach services that direct private-address inputs would block. The same independent re-resolution occurs after manual redirect hops because each redirect target is checked before a separate connection lookup. This issue is fixed in version 4.15.2.
A time-of-check to time-of-use (TOCTOU) race condition in the dynamic loader (ld.so) of the GNU C Library (glibc) versions 2.14 through 2.44 allows a local attacker to escalate privileges. When expanding $ORIGIN in DT_RPATH for setuid/setgid (AT_SECURE) programs, glibc validates the lexically normalized search path against the trusted directories but then opens the raw, un-normalized path. On systems where the Linux fs.protected_hardlinks sysctl is disabled, a local attacker who hard-links such a program into an attacker-controlled directory and wins a race to replace an intermediate path component with a symbolic link can direct the loader outside the trusted directory, causing it to load an attacker-controlled shared object and execute arbitrary code with the elevated privileges of the program.
Exploitation requires an installed setuid or setgid binary whose DT_RPATH uses $ORIGIN followed by ".." traversal that normalizes into a trusted directory, and the ability to hard-link that binary and win the race by swapping a path component for a symbolic link. Major Linux-based OS distributions ship with fs.protected_hardlinks enabled by default and mitigate the vulnerability.
This security update for util-linux addresses multiple vulnerabilities including heap use-after-free, time-of-check to time-of-use (TOCTOU) race conditions, and local privilege escalations affecting components such as libblkid, mount, and nsenter. The fixes improve path canonicalization, restrict unprivileged user operations, and prevent authority leaks and mount helper bypasses. These vulnerabilities could allow local attackers to escalate privileges or interfere with system mount operations.
microsandbox is an easy, fast, local-first microVM runtime and library. Prior to 0.5.10, sdk/rust/lib/runtime/spawn.rs serializes NetworkConfig secret values into the --network-config argument and passes per-sandbox secrets through repeated --env arguments accepted by crates/cli/lib/sandbox_cmd.rs. Other local users or co-resident processes can read these values through the host process table, including /proc process command lines on Linux and process listings on Linux and macOS, for the lifetime of the sandbox. Exploitation does not require code execution inside the sandbox or access to the spawning user's session, and can disclose host-side API keys, tokens, and environment secrets on shared hosts, CI runners, and developer systems. This issue is fixed in version 0.5.10.
Acode is a powerful text and code editor for Android. From 1.11.6 until 1.12.7, com.foxdebug.acode.rk.exec.terminal.TerminalService is declared as an exported service in src/plugins/terminal/plugin.xml without a binding permission, and src/plugins/terminal/src/android/TerminalService.java does not verify the caller. Any installed Android application can bind the service and send MSG_EXEC with an attacker-controlled cmd value, which the terminal implementation passes to ProcessBuilder with sh -c inside Acode's UID. This allows a zero-permission local application to execute commands with access to Acode private data, remote credentials, Storage Access Framework grants, and runtime permissions without additional interaction at attack time. This issue is fixed in version 1.12.7.
PJSIP is a free and open source multimedia communication library written in C. In 2.17 and earlier, the PJSIP AVI parser in pjmedia/src/pjmedia/avi_player.c uses an input-file video chunk length as the number of bytes copied into a frame buffer whose capacity is derived from the declared media dimensions. A crafted AVI file can therefore cause an attacker-controlled out-of-bounds write past the heap allocation when an application plays the file or pulls its frames. The existing size assertion does not protect production release builds, where assertions are disabled. Typical local playback can crash the process, while applications that accept untrusted AVI sources expose a stronger memory-corruption condition. No fixed version is available as of this review.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (46b706e618e75dd4f864a55a33a06cfa98c890d310735b1dea71ad86fe299fe6)
The package ships a module (e.py) that prompts for administrator elevation via UAC, archives the contents of %LOCALAPPDATA%\logs with shutil.make_archive, and POSTs the resulting zip to a hardcoded attacker-controlled endpoint at https://sparkling-pine-f202.stilluer-sweden.workers.dev/ using requests.post. The %LOCALAPPDATA%\logs tree commonly contains Discord, browser, and game-tool artifacts including session data. The package's PyPI summary ("Safe bug reporter for sending local logs via Cloudflare Workers") and identifiers such as aiosendletter_logs are cover-story framing for bulk exfiltration of installer-owned data. Separately, pyproject.toml declares the Python stdlib name "ctypes" as an unpinned PyPI runtime dependency, causing pip to resolve and install whatever package currently occupies that registry name and execute its lifecycle code on the installer's machine — a dependency-confusion install-time code path in addition to the primary stealer payload.
## Source: kam193 (4e81e3b63bbe719f6d0bdd42930bb2f3235b8e41cbacf449743bb7dca67c9f08)
Package hides code to exfiltrate files. Most releases target unclear files, but some reveal the goal to exfiltrate sensitive Telegram data.
---
Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers.
Campaign: 2026-09-aiosendletter
Reasons (based on the campaign):
- files-exfiltration
- target:telegram
A heap out-of-bounds write vulnerability exists in the Linux kernel's cxl/features component due to improper bounds checking of a user-controlled count value against the output buffer size. This flaw allows an attacker with local privileges to cause memory corruption by specifying a count larger than the allocated buffer, potentially leading to system compromise. The issue affects multiple specific Linux kernel versions and has been assigned a high severity with a CVSS score of 7.8.
CVE-2026-90309 is a high-severity vulnerability in the Linux kernel's RDMA/erdma subsystem. It involves improper handling of Completion Queue (CQ) references during Event Queue (EQ) event processing. Specifically, EQ handlers access CQs outside the protection of the xarray lock, which can lead to use-after-free conditions when a CQ is destroyed concurrently. This flaw can cause memory corruption and potentially allow an attacker with local privileges to cause denial of service or escalate privileges.
A vulnerability in the Linux kernel's NTFS driver related to improper handling of kmap_local_page() usage in compressed I/O paths has been resolved. The issue involved discarding the address returned by kmap_local_page() and later accessing or unmapping the page incorrectly, which is invalid for high memory pages. The fix ensures proper mapping and unmapping of pages during decompression and removes unnecessary local mappings in certain write operations.
A memory leak vulnerability in the Linux kernel's FireWire core component was resolved. The issue occurred in the error path of the build_tree() function, where node instances could remain allocated in a local linked list and not be properly released. This could lead to resource leakage when invalid values are detected in the self ID sequence during FireWire device enumeration.
A race condition vulnerability in the Linux kernel's OCFS2 cluster filesystem was resolved by synchronizing heartbeat callbacks with the o2net teardown process. The flaw involved use-after-free errors during concurrent teardown and heartbeat callback execution, potentially causing kernel memory corruption. The patch ensures heartbeat callbacks remain registered until local-node teardown completes, preventing stale or conflicting network state updates.
Authentication Bypass by Spoofing vulnerability in team-alembic ash_authentication allows an attacker who operates one identity-provider connection of a dynamic_oidc strategy to be signed in as a local user established through a different connection.
The strategy is meant to keep each connection in its own identity namespace by writing every UserIdentity row's strategy field as "<name>/<connection_id>", but that namespacing never takes effect. __connection_id__ is populated only on the ephemeral runtime struct built per request in dynamic_oidc/plug.ex, and DynamicOidc.IdentityChange.change/3 re-fetches the strategy from the compile-time DSL through Info.strategy_for_action, yielding the persisted struct whose __connection_id__ is its defstruct default of nil. OAuth2.identity_strategy_name/1 therefore falls back to the bare strategy name for both the identity write and the reads in oauth2/user_resolver.ex and oauth2/sign_in_preparation.ex. Since the identity resource's unique key is (uid, strategy), one row exists per sub across every connection, and the identity-match branch runs before any email check. Neither strategy handles iss, so nothing else distinguishes the issuers: OpenID Connect Core section 5.7 makes sub unique only within an issuer, so two connections numbering subjects independently share one subject space.
This issue affects ash_authentication: from 5.0.0-rc.10 before 5.0.0-rc.14.
A vulnerability in the Linux kernel's NFS server (NFSD) component allows a local administrator with CAP_SYS_ADMIN privileges to crash the kernel by triggering a NULL pointer dereference during certain unlock operations before the server fully starts. This occurs because the code walks NFSv4 client hash tables that have not yet been allocated, leading to a kernel crash. The issue affects specific Linux kernel versions and has been fixed by gating the relevant code paths on a network-up flag to prevent the NULL dereference.
A memory corruption vulnerability (CVE-2026-89161) was found in the PCRE2 library's just-in-time (JIT) compilation function pcre2_jit_match. This flaw can be exploited by a local attacker to cause denial of service or potentially execute arbitrary code. The vulnerability affects Red Hat Hardened Images RPMs containing PCRE2 components. The attack complexity is high and requires local access, reducing overall risk. Red Hat has released updated RPM packages to address this issue.
In Eclipse Ankaios versions 0.1.0 through 1.0.2, the agent creates workload files and Control Interface named pipes (FIFOs) under a predictable path derived from the agent name and a hash of the workload's runtime configuration. If a directory or FIFO already exists at that path when the agent (re)starts, the agent reuses it based only on an existence and/or file-type check, without validating its owner or permissions. A local, unprivileged user with write access to the same base directory (by default under `$TMPDIR/ankaios`, e.g. shared `/tmp`) can pre-create this path hierarchy, including the two Control Interface FIFOs, before the agent starts. The agent then treats the attacker-owned FIFOs as the legitimate Control Interface for the targeted workload. The attacker can complete the Control Interface handshake and issue requests using that workload's configured `controlInterfaceAccess` permissions, allowing impersonation of the workload and, depending on its configured permissions, unauthorized reading and/or modification of the cluster's desired state.
next-video is a library for adding video to Next.js applications. Prior to 2.8.1, the GET endpoint exported by next-video/request-handler and commonly mounted at /api/video accepts an unauthenticated url query parameter, while src/utils/utils.ts isRemote() treats any value without an HTTP or HTTPS prefix as a local path. src/request-handler.ts passes that value through src/assets.ts getAssetPath() to src/config.ts loadAsset(), which appends a JSON suffix and uses fs.readFile without canonicalizing the path or verifying that it remains inside the configured video folder. A remote attacker can therefore escape the intended asset directory and read JSON files accessible to the application process, including Next.js server-action encryption material, preview-mode keys, build manifests, route metadata, absolute paths, and application video asset identifiers. Applications that do not expose the runtime request handler are not affected through this route. This issue is fixed in version 2.8.1.
CVE-2026-90801 is a medium severity buffer overflow vulnerability in GNU Binutils version 2.47. It affects the function cache_bwrite in the ld component, specifically in the file bfd/cache.c. The flaw arises from improper handling of the nbytes argument, leading to a buffer overflow. Exploitation requires local access and no user interaction. Although an exploit has been publicly released, there is no vendor response or patch available at this time.