Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.12.2, `EnvironmentManager.listBackups()` reads each backup's `_manifest.json` and trusts the manifest's `path` field. `EnvironmentManager.pruneBackups()` later passes that trusted `entry.path` directly to `rmSync(entry.path, { recursive: true, force: true })`. An attacker who can place or modify a manifest inside `data/<env>/.backups/<name>/_manifest.json` can cause `network-ai env backup prune --env <env> --keep <n>` or any code path invoking `pruneBackups()` to recursively delete an arbitrary path accessible to the Network-AI process user. This is fixed in v5.12.2. `pruneBackups()` no longer passes `entry.path` from the on-disk manifest to `rmSync`. The deletion path is recomputed from a format-validated `entry.backupId`, and a `dirname` containment check confines deletion to exactly one level under the backups directory. A poisoned manifest (e.g. `"path": "/"`) is now inert.

Network-AI, a TypeScript/Node.js multi-agent orchestrator, has a shipped, exported, documented feature called `ApprovalInbox` (`lib/approval-inbox.ts`). It is the network surface of the human-in-the-loop Approval Gate, which `ApprovalGate` uses to require explicit human approval for high-risk operations. The HTTP server it exposes has no authentication of any kind and sets `Access-Control-Allow-Origin: *` on every route, including the state-changing `POST /approvals/:id/approve` and `/deny`. As a result, in versions 5.0.0 through 5.12.1, any party who can send an HTTP request to the inbox port — a co-located process, a container/SSRF on the same host, a remote client when the operator binds a non-loopback address, or any website the operator visits in a browser (via the wildcard CORS) — can enumerate pending approvals and approve them, defeating the entire human-in-the-loop control and causing the gated high-risk action (e.g. a shell command the agent was holding for review) to exe

Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.12.2, `AgentRuntime` promises scoped file access under a configured sandbox `basePath`, but its path containment checks use raw string prefix tests. A sandbox base such as `/tmp/network-ai-sandbox` also matches a sibling path such as `/tmp/network-ai-sandbox_evil/secret.txt`. An agent/user that can call `AgentRuntime.readFile()` or `AgentRuntime.listDir()` can read or list files outside the intended sandbox when the target path is in a sibling directory sharing the base path prefix. This breaks the documented sandbox boundary. The issue is fixed in v5.12.2. `SandboxPolicy.resolvePath()` and `isPathAllowed()` now use separator-anchored prefix checks (`resolved === base || resolved.startsWith(base + path.sep)`) for both the allow-list and block-list. A sibling directory that merely shares a name prefix (e.g. `/srv/app-evil` vs base `/srv/app`) is no longer treated as in-scope.

Blocksy Companion Pro plugin for WordPress before 2.1.47 contains an unauthenticated arbitrary file upload vulnerability that allows attackers to upload executable files by bypassing extension validation in the save_attachments function exposed through the Advanced Reviews feature. Attackers can exploit the Custom Fonts extension's flawed strpos() substring check by uploading double-extension filenames such as shell.woff2.php, causing the validation to pass on the substring match while the web server executes the file as PHP, achieving remote code execution.

A vulnerability has been found in code-projects Movie Ticketing System 1.0. Impacted is an unknown function of the file /db/moviedb.sql of the component SQL Database Backup File Handler. Such manipulation leads to information disclosure. The attack can be launched remotely. The exploit has been disclosed to the public and may be used.

Sustainable Irrigation Platform (SIP) through version 5.2.16 contains a command injection vulnerability in the optional cli_control plugin that allows unauthenticated or cross-site request forgery attackers to execute arbitrary operating-system commands by storing a malicious payload via the plugin's HTTP endpoint. Attackers can trigger execution by activating the associated irrigation station, exploiting the absence of passphrase protection or the default passphrase 'opendoor', to achieve arbitrary command execution on the underlying host.

Sustainable Irrigation Platform (SIP) through version 5.2.16 contains a server-side request forgery (SSRF) vulnerability that allows unauthenticated attackers to make the device issue arbitrary HTTP requests by supplying a malicious callback URL when the optional Node-RED plugin is installed. Attackers can exploit the lack of destination validation and the default passphrase 'opendoor' to send blind HTTP requests to arbitrary internal or external hosts not otherwise directly accessible.

Sustainable Irrigation Platform (SIP) through version 5.2.16 contains a mass assignment vulnerability that allows unauthenticated attackers to overwrite sensitive configuration settings by supplying arbitrary parameter names in HTTP requests. Attackers can manipulate parameters corresponding to sensitive values such as the passphrase and listening port, and can also achieve the same result through cross-site request forgery due to the absence of adequate request validation.

Sustainable Irrigation Platform (SIP) through version 5.2.16 contains a cross-site request forgery vulnerability that allows remote attackers to perform state-changing administrative actions by luring a logged-in administrator into visiting a malicious page that issues HTTP GET requests without CSRF token validation or origin verification. Attackers can trigger actions such as disabling the passphrase, rebooting the device, deleting programs, or installing plugins, with the default configuration exposing these endpoints to unauthenticated users due to no required passphrase and a default credential of 'opendoor'.

Sustainable Irrigation Platform (SIP) through version 5.2.16 contains a stored cross-site scripting vulnerability that allows unauthenticated attackers to inject arbitrary JavaScript by supplying malicious script payloads within program names submitted via HTTP requests. Attackers can exploit the lack of output encoding on rendered program names to execute arbitrary JavaScript in the browsers of any users viewing the affected page, with exploitation facilitated by the absence of a required passphrase or the default passphrase 'opendoor'.

whichllm before 0.5.16 contains a code injection vulnerability in the run and snippet commands that allows a remote attacker who controls a HuggingFace repository to achieve arbitrary code execution by crafting a malicious GGUF filename containing double quotes or other special characters. The script generation function in cli.py interpolates HuggingFace-derived values, including GGUF variant filenames from the Hub API siblings rfilename field, directly into Python source code without escaping, allowing the crafted filename to break out of the generated string literal and execute injected code on the user's machine before any model download occurs.

Cognee before 1.2.0 contains an improper access control vulnerability that allows unauthenticated attackers to overwrite the global LLM provider configuration by self-registering an account and calling the settings endpoint, which performs no admin or superuser check. Attackers can redirect all LLM operations instance-wide to an attacker-controlled endpoint by exploiting the process-wide singleton configuration cache, enabling exfiltration of prompts, uploaded documents, extracted entities, and knowledge graph content from all users.

GNU Wget through 1.25.0, fixed in commit dd692d9, contains a heap buffer overflow vulnerability in the html_quote_string() function in src/convert.c that allows a remote attacker to trigger memory corruption by supplying a crafted HTML attribute with a large number of characters requiring entity encoding. A server-supplied HTML attribute causes a signed integer counter to overflow during output size accumulation, resulting in an undersized heap allocation and subsequent heap buffer overflow during the copy phase.

GNU Wget through 1.25.0, fixed in commit c2640fe, contains a heap buffer overflow vulnerability in the convert_fname() function within src/url.c that allows remote attackers to trigger memory corruption through a server-supplied filename requiring character set conversion. When the output buffer is too small during iconv E2BIG reallocation, the reallocation logic miscalculates the remaining space, leading to a heap buffer overflow that can be exploited via a maliciously crafted server response.

GNU Wget through 1.25.0, fixed in commit 43d3ba9, contains an integer overflow vulnerability in the parse_content_range() function within src/http.c that allows server-controlled values to cause signed integer arithmetic to overflow. Attackers can supply malicious Content-Range header values to trigger undefined behavior and download desynchronization in the affected client.

The affected Watchfire Controller Software contains self-signed hard-coded RSA private keys and corresponding X.509 certificates used for authenticating and encrypting HTTPS/TLS connections to the controller's built-in web management interface. These keys are embedded in plaintext within the application patch binaries in the firmware directly from Watchfire's Remote Support filestore.

GNU Wget through 1.25.0, fixed in commit 37a40fc, contains a heap buffer underread vulnerability in the clean_metalink_string() function within src/metalink.c that allows a malicious server to trigger memory corruption by serving a Metalink document containing a whitespace-only URL. Attackers can cause the function to decrement a pointer past the start of the buffer when processing an all-whitespace Metalink URL, potentially leading to abnormal program behavior.

NocoBase through 2.1.20 contains a server-side request forgery vulnerability in the serverRequest wrapper that allows authenticated administrators to issue arbitrary outbound HTTP requests by supplying malicious URLs to workflow request nodes, custom request action buttons, or the AI plugin. Attackers can target loopback addresses, RFC-1918 private ranges, and cloud instance metadata endpoints to perform internal network port enumeration, host discovery, and retrieval of IAM role credentials from the instance metadata service. v2.1.18 added a warning message for when SERVER_REQUEST_WHITELIST is not configured.

Cockpit CMS through 2.14.0 contains a path traversal and local file inclusion vulnerability that allows unauthenticated attackers to read arbitrary files or execute PHP files by including unvalidated PATH_INFO derived from REQUEST_URI in filesystem path construction without containment checks. Attackers can inject dot-dot sequences into the URL to traverse outside the designated spaces directory, and when the resolved path ends with a .php extension, the application passes it to include(), enabling local file inclusion on deployments using the PHP built-in server or certain non-default Nginx configurations.

AutoBangumi before 3.2.8 contains a hard-coded default credentials vulnerability that allows unauthenticated attackers to authenticate as the administrator by using the publicly known default credentials seeded at startup via add_default_user() in the database user module when the users table is empty. Attackers can submit the default credentials to the authentication login endpoint to gain full control of the application, including RSS feed configuration, downloader configuration, and all authenticated API endpoints.