A vulnerability in the Linux kernel's fuse direct write implementation causes incorrect invalidation of file data ranges after O_APPEND writes. This leads to potential data inconsistency when file sizes are not page-aligned, as stale data may be read by concurrent readers. The issue is due to capturing the file position before it is updated to EOF, resulting in invalidation of a stale range rather than the actual written range.
A use-after-free vulnerability in the Linux kernel's md/md-llbitmap component was resolved. The issue occurs when the llbitmap_create() function publishes a bitmap pointer before fully reading the bitmap superblock, leading to potential stale pointer access by readers such as /proc/mdstat. The fix ensures that the bitmap pointer is cleared while holding the appropriate mutex, preventing readers from seeing a freed bitmap.
A vulnerability in the Linux kernel's batman-adv module allowed non-atomic updates of MAC addresses, potentially causing readers to see partially updated MAC addresses. This could lead to network issues such as ARP cache poisoning. The issue has been fixed by using atomic64_t to ensure atomic updates of the 48-bit MAC address, preventing partial reads.
A synchronization issue in the Linux kernel's bcm-vk driver was addressed by properly using acquire/release semantics for the msgq_inited flag. The flaw allowed readers to observe msgq_inited set while still seeing stale message queue information due to lack of proper memory ordering with atomic operations.
Use of a One-Way Hash with a Predictable Salt vulnerability in team-alembic AshAuthentication allows readers of the audit store to recover the client IP addresses that the audit log add-on's :hash privacy mode is meant to pseudonymise.
AshAuthentication.AddOn.AuditLog.IpPrivacy.hash_ip/1 computes a single unkeyed :crypto.hash(:sha256, salt <> ip) and truncates the result to 16 hexadecimal characters. The salt is read from the :audit_log_ip_salt or :secret application config keys, and falls back to the constant "default-salt-change-in-production" published in the library source when neither is set, with nothing warning that the default is in use. The IPv4 space is only 2^32 values and SHA-256 is fast, so the whole hash table is precomputable and every stored value maps back to its source address. Truncating to 16 characters does not help, and even a configured salt leaves the hash cheap enough to enumerate once it leaks.
This issue affects ash_authentication: from 4.12.0 before 4.15.0 and from 5.0.0-rc.0 before 5.0.0-rc.14.
Insertion of Sensitive Information into Log File vulnerability in team-alembic AshAuthentication allows disclosure of user password digests to readers of the audit store.
The audit_log add-on builds each entry's extra_data in AshAuthentication.AddOn.AuditLog.Auditor.build_extra_data/4, which takes :actor from the action callback context verbatim. Any audited action invoked with actor: set to a user record therefore deposits that record, including its hashed_password attribute, into the audit entry. The same module already collapses the audited identity to an opaque string via AshAuthentication.user_to_subject/1 and filters params against the strategy's configured allow-list, so the actor is the only stored value that reaches the audit store unfiltered. Marking the attribute sensitive?: true does not help, because that redacts inspect/1 output rather than JSON encoding or raw-term storage.
There is no attacker-controlled trigger and no network disclosure path: entries accumulate from ordinary authenticated activity, and an attacker's own requests deposit only their own digest. Exploitation requires independent read access to the audit store, such as database credentials, an audit role, a backup, or a log shipper, at which point the digests support offline password attack against every active account. Whether the material persists depends on the data layer, since raw-term stores keep it verbatim while a SQL store raises Protocol.UndefinedError and drops the entry unless the user resource derives Jason.Encoder.
This issue affects ash_authentication: from 4.12.0 before 4.15.0 and from 5.0.0-rc.0 before 5.0.0-rc.2.
Wire provides gRPC and protocol buffers for Android, Kotlin, Swift, and Java. Prior to 6.4.5 and 7.0.0-alpha04, Wire protobuf readers do not consistently validate attacker-controlled lengths against the current logical message boundary before advancing cursors, pointers, limits, slices, or allocations. In Kotlin, ProtoAdapter.decode(ByteArray) and ProtoAdapter.decode(ByteString) use ByteArrayProtoReader32.internalNextLengthDelimited(), where a positive oversized length can wrap pos + length to a negative limit and escape the existing negative-length check. Related ProtoReader, ReadBuffer.readVarint(), ReadBuffer.verifyAdditional(count:), packed-repeated, nested-message, and ProtoDecoder.decodeSizeDelimited(_:from:) paths can cross logical boundaries, perform pointer arithmetic, reserve capacity, or convert an unrepresentable size before proving the requested bytes exist. An attacker who supplies malformed protobuf bytes can cause unchecked exceptions, traps, out-of-bounds behavior, or excessive allocation, resulting in denial of service without known confidentiality, integrity, or code-execution impact. This issue is fixed in versions 6.4.5 and 7.0.0-alpha04.
CVE-2026-89972 is a critical vulnerability in the Linux kernel's NVMe subsystem where a missing SRCU grace period in an error path can lead to use-after-free conditions. Specifically, the error path in nvme_alloc_ns() removes a namespace from a list without waiting for concurrent readers to finish, potentially causing a freed namespace structure to be accessed. This flaw was fixed by adding the necessary SRCU synchronization to ensure safe memory reclamation.
A use-after-free vulnerability in the Linux kernel's sunrpc subsystem was resolved by deferring the freeing of rq_argp and rq_resp until after the RCU grace period. The flaw allowed RCU readers to access freed memory, potentially causing system instability or security issues. The fix moves the freeing of these resources into an explicit call_rcu() callback alongside the struct free.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (11eb7019f9ead481b3c5f203703df3b8b63abab2e03502a02ba30360da9948d9)
The `ollamaa` package on PyPI is a one-character typosquat of `ollama`. Its `__init__.py` re-exports everything from the real `ollama` package so imports appear to function normally, while the tarball installs a `.pth` file into site-packages. Python auto-executes lines beginning with `import` in every installed `.pth` on every interpreter startup, so this file runs without any explicit import of the package. The `.pth` opens with cover-story comments instructing readers and automated analyzers to classify the file as clean, then invokes `exec` on a payload obfuscated by base64 followed by XOR with 0x5A, gated by a `/tmp/.lurves-planted` first-run sentinel. The decoded payload spawns a detached subprocess that fetches `http://167.86.108.190:7788/.lurves-agent.py` over plaintext HTTP from a bare IP, writes it to `/tmp/.a`, and runs `python3 /tmp/.a --daemon` with `start_new_session=True` and suppressed output. The result is unauthenticated full-host remote code execution triggered on the next Python invocation after install, with attacker-supplied code fetched at runtime from a hardcoded bare-IP endpoint.
## Source: kam193 (bdddd369c97996ed3c886729053c3f8189bf96bc22914f2c61c901b3e96073fc)
Typosquatting package planting a malicious PTH file that during loading Python downloads the next stage. It then hides its existence, downloads the further stage and clears logs. The next stage exfiltrates sensitive data (SSH keys, cloud credentials), ensures persistence and plants a cryptominer. The original PTH file contains a simple attempt to discourage analysis via AI agents.
---
Category: MALICIOUS - The campaign has clearly malicious intent, like infostealers.
Campaign: 2026-09-openaii
Reasons (based on the campaign):
- Downloads and executes a remote malicious script.
- obfuscation
- abuses-pth
- cryptominer
- infostealer
- exfiltration-credentials
- files-exfiltration
- exfiltration-ssh-keys
- persistence
- typosquatting
- covering-tracks
A race condition vulnerability in the Linux kernel's ring buffer implementation (trace_buffer) has been fixed. The issue involved unsafe lockless reads of the sub-buffer size during concurrent resizing, which could lead to inconsistent or corrupted data handling. The fix involves using safer accessors and synchronizing buffer modifications with a mutex to prevent race conditions.
CVE-2026-89747 is a use-after-free vulnerability in the Linux kernel's tracing subsystem affecting trace_pipe reads when the sub-buffer order changes. The flaw occurs because the ring buffer sub-buffers are freed and replaced while readers still hold pointers to them, leading to dereferencing freed memory. This can cause kernel crashes or potentially allow privilege escalation. The issue is fixed by taking a lock around the order change to prevent races with readers.
CVE-2026-89708 is a critical use-after-free vulnerability in the Linux kernel's NFS server (nfsd) related to the handling of callback sessions during session teardown. The flaw occurs because the session memory can be freed while an inflight RPC callback task still holds a reference to it, leading to potential dereferencing of freed memory. The issue is fixed by converting the session pointer to be RCU-protected, ensuring safe memory reclamation after all readers have finished accessing it.