Nedlir

An unsanitized deserialization issue allows authenticated attackers to manipulate patient-matching criteria. By providing a crafted serialized `key` parameter to the 'interface/orders/patient match dialog.php' endpoint, attackers can enumerate patients or force matches that associate lab orders with the incorrect patient. Deserialization is the process of converting a serialized string back into an object or data structure.

A security scan bypass allows attackers to hide malicious payloads. This occurs because the functions `collect bundle siblings()` and `walk bundle subdir()` skip nested directories and .pyc files (compiled Python bytecode files). An attacker can distribute a skill bundle containing a .pyc file with an unchecked hash located under `scripts/<subdir>/ pycache /`, which the system ignores during scanning but CPython executes during runtime.

A flaw in the Static Scanner component allows a local attacker to cause a protection mechanism failure. This issue occurs within the `shouldSkipTargetPath()` function located in the internal/runner/runner.go file.

A path traversal issue exists within the File Access component, specifically affecting the `startsWith()` function in the `skill scan/tools/dir/dir actions.py` file. This flaw allows a local attacker to manipulate file paths to access unauthorized directories.

The `isSecurityDefenseSkill()` function within the `dist/scanner/analyzers/context.js` file relies on untrusted inputs when making security decisions. This issue allows a local attacker to perform a manipulation that compromises the security decision process.

A flaw in the Static Scanner component, specifically within the file `internal/runner/static scanner.go`, allows for an incomplete comparison with missing factors. This issue can be exploited by an attacker on the local host.

An interpretation conflict exists within the File Classifier component. The issue resides in the `IsBinaryFile()` function located in the internal/runner/static scanner.go file. Exploitation requires local access to the system.

A denial of service can be triggered remotely via the xAI OAuth Callback Handler component. The issue exists in the `waitForCallback()` function within the `src/services/api/xaiOAuthCallback.ts` file, where manipulation of the `Error` argument allows for the attack.

The software fails to analyze executable files when scanning local skill directories. It only reads the SKILL.md file and ignores Python source code, bytecode, and other artifacts located in the scripts directory. This allows malicious code to be distributed within non-manifest files, which then receive a SAFE verdict and a 100/100 trust score because the executable payloads remain unanalyzed.

The software fails to analyze compiled Python bytecode files located in companion code directories. This allows attackers to bypass security scanning by including malicious ` pycache ` entries alongside benign source files. Consequently, arbitrary Python bytecode can be executed upon import, even when the scanner provides a CERTIFIED verdict with high trust scores in both static and semantic analysis modes.

An unauthenticated caller can repeatedly invoke the `sendForgotPasswordEmail` Meteor method because it lacks a DDP rate limit. This method is accessible via DDP and the HTTP endpoint 'POST /api/v1/method.callAnon/sendForgotPasswordEmail'. Exploitation allows an attacker to send an unlimited volume of password reset emails to a specific address using the system's mail sender. Additionally, the method can be used to probe for accounts; it returns true for addresses without an account or for successful sends, but returns false for accounts using external providers when `Accounts AllowPasswordChangeForOAuthUsers` is disabled, allowing an attacker to identify those specific account types.

Insufficient access control granularity occurs when the `CONTAINERS` environment variable is set, failing to properly gate read endpoints within the `/containers` Docker API namespace. This allows attackers to use GET requests to the endpoints '/containers/{id}/archive', '/containers/{id}/export', '/containers/{id}/logs', and '/containers/{id}/top' to read arbitrary files and download entire container filesystems as tar archives.

An unrestricted upload issue exists within the Document Upload Controller component. A remote attacker can exploit the `DocumentController()` function located in the `grails-app/controllers/org/pih/warehouse/core/DocumentController.groovy` file to upload files without proper restrictions.

A remote manipulation of the Product Supplier Edit Controller within the file `grails-app/controllers/org/pih/warehouse/RoleInterceptor.groovy` can lead to improper authorization.

Remote manipulation of the `needManager()` function within the Role Interceptor component (located in grails-app/controllers/org/pih/warehouse/RoleInterceptor.groovy) can lead to improper privilege management.

A server-side request forgery (SSRF) exists in the Product Upload Endpoint. A remote attacker can manipulate the `params.url` variable within the `Upload()` function of the file grails-app/controllers/org/pih/warehouse/product/ProductController.groovy to trigger the issue. SSRF is a flaw that allows an attacker to induce the server-side application to make requests to an unintended location.

An issue exists in the Template Processing component where the `buildZebraTemplate()` function in the `grails-app/controllers/org/pih/warehouse/core/DocumentController.groovy` file fails to properly neutralize special elements used in a template engine. This allows a remote attacker to manipulate the template processing.

The skill-scan component fails to analyze compiled Python bytecode files because it hardcodes ` pycache ` directories and `.pyc`, `.pyo`, and `.pyd` extensions into skip lists. This allows an attacker to bundle malicious compiled bytecode with benign Python source files. Since the scanner ignores the bytecode, it may return a safe verdict, allowing the execution of arbitrary code when an operator installs the skill.

A remote resource consumption issue exists within the GraphQL API Endpoint component, specifically affecting the library `packages/core/src/lib/core/queries/output-field.ts`. This allows a remote attacker to manipulate the system, leading to excessive resource usage.

The JSONSerializer and CBORSerializer are subject to Remote Code Execution (RCE) through insecure deserialization. The `unmarshal object()` function enables arbitrary class instantiation and state injection by dynamically importing modules and calling ` setstate ` on any class present in the Python environment. This can be triggered by submitting a specially crafted JSON or CBOR payload to an application utilizing these serializers.