Qerogram

Pi loads project-local configuration and resources from a repository's `.pi` directory without requiring user trust. This includes project-local extensions, which are executable TypeScript or JavaScript modules loaded directly into the Pi process. An attacker controlling a repository can place malicious Pi-specific resources within it; if a user starts Pi from that working tree, the extension code executes with the same privileges as the local Pi process. The affected component is the project resource loading path, which previously initialized `.pi/settings.json`, project package-managed resources, and project instruction files during session initialization before a trust boundary was established. Because extensions are not sandboxed, they can register commands, tools, and event handlers, allowing access to files, environment variables, credentials, the network, and local tools available to the user.

Cursor is a code editor built for programming with AI. A UI information disclosure exists in Cursor's MCP (Model Context Protocol) deeplink handler, enabling attackers to execute arbitrary system commands through social engineering attacks. Clicking malicious `cursor://anysphere.cursor-deeplink/mcp/install` links does not display the command arguments in the installation dialog. If a user clicks a malicious deeplink and proceeds with the installation, the full command, including arguments, will be executed on the machine.

The issue allows arbitrary data to be written to arbitrary files, leading to a Remote Code Execution.

The issue concerns an import functionality in the HTTP Headers WordPress plugin that executes arbitrary SQL on the server, leading to an SQL Injection vulnerability. This allows for potential exploitation by executing unauthorized SQL commands.

The issue is related to a SQL injection that occurs due to improper sanitization and escaping of a parameter in a SQL statement when deleting messages. This can be exploited by high-privilege users, such as administrators.

The issue allows high privilege users, such as admins, to perform Stored Cross-Site Scripting attacks. This can occur even when the unfiltered html capability is disallowed, for example in a multisite setup.

The issue allows a high-privileged user, such as an admin, to upload a zip file containing malicious PHP code, leading to remote code execution. This is due to the plugin not properly validating uploaded custom icon packs.

The issue allows high privileged users to inject arbitrary javascript, even when the unfiltered html capability is disallowed, due to improper escaping of the `image file` field when adding a new social icon.

The issue allows attackers to execute arbitrary code via a crafted HTML file, exploiting an arbitrary file write vulnerability.

The issue arises from the lack of sanitization of the `selected icons` attribute in the `cnss widget` before it is used in an SQL statement, leading to a SQL injection vulnerability.

The issue allows a high privileged user to bypass certain security settings and upload arbitrary files to the site through the `ajax save` function. The file is written relative to the current stylesheet directory with a .php file extension added, and no validation is performed on the file's content. This can trigger a remote code execution (RCE) vulnerability by uploading a web shell. Additionally, the `name` parameter is not sanitized, allowing the payload to be uploaded to any directory the server has write access to.

The issue allows high privilege users, such as admins, to query arbitrary files on the web server via a path traversal vector due to improper validation of the path of files to be searched. Additionally, a flaw in the search function allows a pattern to be provided, which can be used to output relevant matches from the matching file, potentially disclosing all content of the file.

The issue allows logged-in users with the custom "Amelia Manager" role to potentially upload PHP backdoors onto the site. This is because the plugin stores image blobs into actual files whose extension is controlled by the user.

The issue allows high privilege users to perform SQL injection attacks due to the lack of sanitization and escaping of the `rm form id` parameter in the Automation admin dashboard.

The issue allows high privilege admins to upload arbitrary PHP files, potentially leading to remote code execution (RCE), even when the blog is hardened with constants such as DISALLOW UNFILTERED HTML, DISALLOW FILE EDIT, and DISALLOW FILE MODS set to true.