CVE Feed

An authorization issue was addressed with improved entitlement checks. This issue is fixed in macOS Sequoia 15.8, macOS Tahoe 26.7. An app with admin privileges may be able to modify the System Keychain.

An input validation issue was addressed with improved input validation. This issue is fixed in Xcode 27. Parsing a maliciously crafted binary may lead to unexpected app termination.

An out-of-bounds read was addressed with improved input validation. This issue is fixed in Xcode 27. Parsing a maliciously crafted binary may lead to unexpected tool termination.

A logic issue was addressed with improved state management. This issue is fixed in Safari 27, visionOS 27, watchOS 27. Processing maliciously crafted web content may lead to unexpected behavior.

An integer underflow was addressed with improved input validation. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. An attacker with control of a network file server may be able to cause unexpected system termination.

An authentication issue was addressed with improved state management. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. An attacker may be able to impersonate other users.

An out-of-bounds write issue was addressed with improved bounds checking. This issue is fixed in iOS 26.7 and iPadOS 26.7, iOS 27 and iPadOS 27, macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7, tvOS 27, visionOS 27, watchOS 27. An attacker in radio range may be able to cause unexpected system termination.

A privacy issue was addressed with improved access restrictions. This issue is fixed in iOS 27 and iPadOS 27, macOS Golden Gate 27, tvOS 27, visionOS 27, watchOS 27. An app may be able to access sensitive user data.

A logic issue was addressed with improved checks. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. Opening a file referencing a network share may lead to unexpected software execution.

A memory corruption issue was addressed with improved input validation. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. An attacker in physical proximity may be able to corrupt process memory.

This issue was addressed with improved validation of the caller's access. This issue is fixed in iOS 27 and iPadOS 27, macOS Golden Gate 27, macOS Tahoe 26.7, visionOS 27. An app may be able to access protected user data.

This issue was addressed through improved state management. This issue is fixed in iOS 27 and iPadOS 27. An attacker with physical access to a locked device may be able to bypass Digital ID authentication when iPhone Mirroring is active.

A buffer overflow was addressed with improved size validation. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. An app may be able to cause unexpected system termination.

The issue was addressed with improved handling of secure input state. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. An app may be able to access sensitive user data.

A race condition was addressed with additional validation. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. A malicious application may be able to bypass Privacy preferences.

An access issue was addressed with improved access controls. This issue is fixed in macOS Golden Gate 27, macOS Sequoia 15.8, macOS Tahoe 26.7. An app may be able to access sensitive user data.

The issue was addressed with additional app bundle integrity checks. This issue is fixed in macOS Golden Gate 27. A maliciously modified app may not be prevented from launching.

The issue was addressed with improved checks. This issue is fixed in macOS Sequoia 15.8, macOS Tahoe 26.7. An app may be able to elevate privileges.

An authorization issue was addressed with improved state management. This issue is fixed in macOS Sequoia 15.8, visionOS 27. An app may be able to access sensitive user data.

Summary Before version 0.9.0, Pingora accepted HTTP/2 requests whose Host header differed from the :authority pseudo-header. Request filters, cache keys, and logs that read the request URI (session.req_header().uri) saw the :authority value, but the upstream received the client-supplied Host value. Pingora also accepted :authority values containing userinfo (for example, [email protected]). This violates RFC 9113 Section 8.3.1, which requires an intermediary to use the :authority value for Host and replace any existing Host field. Root cause * HTTP/2 to HTTP/1.1: Host was generated from :authority only when the request had no Host header, so a conflicting client Host was forwarded unchanged. * HTTP/2 to HTTP/2: the client Host header was removed and used to overwrite the upstream :authority, after request filters had already run against the original :authority. * The HTTP/2 server did not compare Host with :authority and did not reject userinfo in the authority. Impact A remote, unauthenticated attacker can send a single HTTP/2 request that the application sees as one host and the upstream receives as another. Where request filters, access controls, or cache keys use the URI authority and the upstream routes by Host, this can lead to: * Cache poisoning: a response from another virtual host is stored under the victim host's cache key and served to other users (this also requires the upstream to serve different content per Host). * Host-based access control bypass: a request approved for one host is served by a different virtual host on the upstream. Deployments that do not accept HTTP/2 from clients, or that do not use the request authority for security or caching decisions, are not affected. Note that Cloudflare CDN customers were unaffected, as internal Pingora services at Cloudflare do not have HTTP/2 clients. Mitigation: * Upgrade to pingora 0.9.0 or later. Pingora now returns 400 Bad Request for HTTP/2 requests whose Host differs from :authority, that contain more than one Host, or whose authority contains userinfo, and keeps the upstream Host or :authority consistent with the request authority. Fix commits: https://github.com/cloudflare/pingora/commit/915590a91811877795531f6ba0927a3138b5cb4b , https://github.com/cloudflare/pingora/commit/ffab8302d08d4292d1924eb8fe84647fd5cb8e76 * If you cannot upgrade: in request_filter, reject HTTP/2 requests that carry a Host header different from the URI authority, or whose authority contains '@'.

Letta through 0.16.8 contains an information disclosure vulnerability that allows users running tools to read server secrets because LocalSandbox subprocesses inherit the full server environment. Attackers can run code via POST /v1/tools/run or custom tools to read LETTA_PG_URI, POSTGRES_PASSWORD, and provider API keys, then access cross-tenant database data.

Letta through 0.16.8 contains an authentication bypass vulnerability that allows API callers to impersonate other users by supplying a self-asserted user_id HTTP header. Attackers can set user_id to another tenant's ID, even with LETTA_SERVER_SECURE enabled, to read and modify that user's agents, memory and configuration.

Letta through 0.16.8 ships a compose.yaml that does not enable LETTA_SERVER_SECURE, so the password middleware is never added and all API routes are unauthenticated, and it exposes a superuser PostgreSQL port with default credentials. Remote attackers can execute arbitrary Python via POST /v1/tools/run, manage /v1/admin users, or run OS commands via COPY PROGRAM.

Coze Studio through 0.5.1 contains an insecure direct object reference vulnerability in workflow.go that allows authenticated users to access other spaces' workflows by supplying foreign workflowIDs. Attackers can pair their own spaceID with another tenant's workflowID in save, update_meta, batch_delete, canvas, test_run and nodeDebug routes to read, run, overwrite or delete workflows.

Coze Studio before 0.5.0 contains an insecure default configuration that allows any registered user to execute commands by running workflow code nodes in an unsandboxed python3 subprocess. Attackers can bypass the regex import filter with __import__('subprocess') to run commands, stealing environment variables, model API keys, and accessing MySQL, MinIO, Redis and etcd.

DB-GPT through 0.8.2 contains a hardcoded credentials vulnerability in its default docker-compose.yml that sets a static MYSQL_ROOT_PASSWORD and publishes port 3306 on all interfaces. Attackers reaching port 3306 or the shared dbgptnet network can authenticate as root@% to gain full database access including FILE and SUPER privileges.

DB-GPT through 0.8.2 contains an unauthenticated remote code execution vulnerability because the /api/v1/chat/react-agent endpoint relies on a mock get_user_from_headers function that grants every caller the admin role. Unauthenticated attackers can send prompts that make the ReAct agent's unsandboxed code_interpreter and shell_interpreter tools run arbitrary Python or bash as root in the webserver container, including reaching the bundled database with default credentials.

EvoScientist 0.1.0 before 0.2.5 contains an approval bypass vulnerability that allows attackers to run host shell commands by delegating them to async sub-agents lacking an interrupt_on gate. Attackers can plant instructions in ingested content such as fetched web pages, so the agent calls start_async_task and runs commands with user privileges and environment secrets.

Aix-DB 1.2.4 contains a remote code execution vulnerability that allows authenticated non-admin users to run operating system commands through the ChatBI common agent's ungated deepagents execute tool. Attackers can send crafted chat input to POST /dify/get_answer or upload documents with malicious text to run host commands, exposing datasource credentials and model API keys.

Langchain-Chatchat through 0.3.1.3 exposes a POST /tools/call API endpoint that invokes any registered tool by name without authentication, leaving the always-registered shell tool callable. Remote attackers can send a single request naming the shell tool to execute arbitrary OS commands as the server user (root in the shipped Docker image).

OpenSRE through 0.1.2026.10.19 contains an approval gate bypass vulnerability that allows attackers to execute shell_run commands without confirmation because the headless presenter's execution_allowed() always returns True. Attackers can inject instructions into alerts or Slack messages triaged by the agent to run commands on the host and exfiltrate LLM, Slack and cloud credentials.

JoySafeter through 1.0.0 contains a remote code execution vulnerability that allows authenticated users to run operating system commands via an unsandboxed execute tool in Copilot standard mode. Attackers can self-register, select standard mode in the WebSocket chat.start extension, and have the agent invoke subprocess.run with shell=True on the host to access other tenants' data and server secrets.

piflow 2.0.0 through 2.0.1 contains a remote code execution vulnerability in its DeepAgents chat agent, whose POST /chat and POST /chat/stream endpoints lack authentication. Attackers can instruct the agent to call the exec_shell tool, which runs arbitrary commands via subprocess.run without human approval or an allow-list, exposing environment secrets such as LLM API keys and database credentials.

LangAlpha through 2026.09.07 contains an improper isolation vulnerability that places all tenants' sandbox containers on Docker's shared default bridge network with inter-container communication enabled. Authenticated users with code execution in their own sandbox can connect directly to other tenants' containers, bypassing require_workspace_owner checks to read files such as vault secrets from exposed servers.

LangAlpha through 2026.09.07 contains an indirect prompt injection vulnerability that allows remote attackers to execute shell commands or Python by planting instructions in web content retrieved by the agent. Because Bash and ExecuteCode lack an allow-list and human-in-the-loop approval, attackers can run code in the workspace sandbox and exfiltrate vault secrets from _internal/.vault_secrets.json.

Ning compress-lzf before 1.2.0 contains an out-of-bounds memory access vulnerability in the sun.misc.Unsafe based UnsafeChunkDecoder that fails to validate back-references and literal runs. Attackers can supply crafted LZF streams to applications using the default LZFDecoder to leak unrelated JVM heap memory into decompressed output or crash the JVM.

Summary Before version 0.8.1, Pingora's HTTP/2 server used the h2 crate's default limits unless the application set its own 16 MiB decoded header-list size and no limit on concurrent streams. A remote, unauthenticated client can send 1-byte HPACK references to a near-empty header, making the server allocate hundreds of thousands of header fields per stream across an unlimited number of streams on one connection, and use HTTP/2 flow control to keep those streams open. This can exhaust memory and degrade or terminate the Pingora process. Impact Remote, unauthenticated denial of service against any Pingora service that terminates HTTP/2 (h2 or h2c) without explicitly setting max_header_list_size and max_concurrent_streams. The reporter showed that a single connection could get a Pingora process with a 2 GiB memory limit OOM-killed. Note that Cloudflare CDN customers were unaffected, as internal Pingora services at Cloudflare do not have HTTP/2 clients. Mitigation: * Upgrade to pingora 0.8.1 or later (fix: https://github.com/cloudflare/pingora/commit/d193c8d49b8b7c1c1ede93183759caa4f6906bbd ). When no H2Options are supplied, Pingora now advertises and enforces a 64 KiB decoded header-list limit and at most 100 concurrent streams per connection. * Applications that supply their own H2Options are not covered by these defaults and must set max_header_list_size and max_concurrent_streams explicitly (this also works on versions before 0.8.1), or, on 0.8.1 and later, start from pingora_core::protocols::http::v2::server::default_h2_options(). H2Options::default() does not include these limits.

A flaw was found in the iscsiuio component of open-iscsi. The DHCPv6 option parser calls strlen() on raw network data without ensuring a NUL terminator is present, causing it to read past the buffer into heap memory. This can result in a daemon crash, a 1-byte heap overwrite, or corruption of the iSCSI initiator name. An attacker on the same network segment can trigger this by sending a crafted DHCPv6 Advertise message.

A flaw was found in the iscsiuio component of open-iscsi. The IPv6 Neighbor Discovery handler accesses link-layer address options at a fixed offset without verifying the packet contains them. Additionally, the ARP table is updated unconditionally with attacker-controlled addresses from the packet headers. An attacker on the same network segment can send a minimal Neighbor Advertisement to poison the ARP table and redirect network traffic.

A flaw was found in the iscsiuio component of open-iscsi. The DHCPv4 option parser incorrectly calculates the length of the options data, causing it to read beyond the allocated buffer into adjacent heap memory. An attacker on the same network segment can send a crafted DHCP response to crash the daemon or corrupt its network configuration.

A segmentation violation in the gf_props_equal_internal function (/filter_core/filter_props.c) of GPAC Project/MP4Box before 2.5 allows attackers to cause a Denial of Service (DoS) via supplying a crafted data file.

A Floating-Point Exception (FPE) in the reframer_send_packet function (/filters/reframer.c) of GPAC Project/MP4Box before 2.5 allows attackers to cause a Denial of Service (DoS) via supplying a crafted MP4 file.

An issue in nvbn thefuck v.3.32 allows a local attacker to execute arbitrary code via the get_rules_import_paths in thefuck/corrector.py and Rule.from_path in thefuck/types.py

A buffer overflow in the gf_ac4_parser_bs function (/media_tools/av_parsers.c) of GPAC Project/MP4Box before 2.5 allows attackers to cause a Denial of Service (DoS) via supplying a crafted input.

IBM Sterling Connect Direct for Unix 6.3.0.0 through 6.3.0.7 and 6.4.0.0 through 6.4.0.6 could allow a local attacker to gain elevated privileges due to improper privilege management.

IBM Langflow OSS 1.0.0 through 1.12.4 contains three message endpoints that perform no ownership verification. Any authenticated user can bulk-delete another user's messages, overwrite message content to inject malicious prompts, or hijack an entire conversation session by reassigning its session ID.

IBM TRIRIGA Application Platform 5.0.0 through 5.0.8 could allow a remote authenticated attacker to execute arbitrary code due to improper authentication.

IBM Verify Identity Access 11.0.3.1 and IBM Security Verify Access 10.0.9.3 Appliance could allow a physically proximate attacker to gain elevated privileges due to the use of hard-coded credentials.

IBM ContextForge MCP Gateway 0.1.0 through 1.0.5 is vulnerable to a server-side request Forgery (SSRF) that allows authenticated users with gateway creation privileges to perform arbitrary HTTP POST requests to internal network resources and cloud metadata endpoints. The vulnerability occurs during OAuth-enabled gateway registration when the system fails to validate the oauth_config.token_url parameter. When registering a gateway with OAuth client credentials authentication, an attacker can supply a malicious token_url pointing to internal services, cloud metadata endpoints (AWS IMDSv1, GCP, Azure), or localhost-bound services. The application immediately performs an HTTP POST request to the attacker-controlled URL with form-encoded credentials in the body. If the response cannot be parsed as a valid OAuth token response, up to 256 characters of the response body are reflected back to the attacker in the HTTP 502 error message. This vulnerability bypasses existing SSRF protections that are applied to the gateway.url field but not to OAuth configuration parameters.

A security vulnerability has been detected in H3C Magic NX15 NX15V100R017. This impacts the function sub_416874 of the file /usr/bin/usrlist. Such manipulation leads to stack-based buffer overflow. It is possible to launch the attack remotely. The exploit has been disclosed publicly and may be used. The vendor was contacted early about this disclosure.