Weakness (CWE)
CWE-94: Improper Control of Generation of Code
The product constructs all or part of a code segment using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the syntax or behavior of the intended code segment.
How to prevent it (7 mitigations)
-
Architecture and Design
Refactor your program so that you do not have to dynamically generate code.
-
Architecture and Design
Run your code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which code can be executed by your product.
Examples include the Unix chroot jail and AppArmor. In general, managed code may provide some protection.
This may not be a feasible solution, and it only limits the impact to the operating system; the rest of your application may still be subject to compromise.
Be careful to avoid CWE-243 and other weaknesses related to jails.
-
Implementation
Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does.
When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue."
Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright.
To reduce the likelihood of code injection, use stringent allowlists that limit which constructs are allowed. If you are dynamically constructing code that invokes a function, then verifying that the input is alphanumeric might be insufficient. An attacker might still be able to reference a dangerous function that you did not intend to allow, such as system(), exec(), or exit().
-
Testing
Use dynamic tools and techniques that interact with the product using large test suites with many diverse inputs, such as fuzz testing (fuzzing), robustness testing, and fault injection. The product's operation may slow down, but it should not become unstable, crash, or generate incorrect results.
-
Operation
Run the code in an environment that performs automatic taint propagation and prevents any command execution that uses tainted variables, such as Perl's "-T" switch. This will force the program to perform validation steps that remove the taint, although you must be careful to correctly validate your inputs so that you do not accidentally mark dangerous inputs as untainted (see CWE-183 and CWE-184).
-
Operation
Run the code in an environment that performs automatic taint propagation and prevents any command execution that uses tainted variables, such as Perl's "-T" switch. This will force the program to perform validation steps that remove the taint, although you must be careful to correctly validate your inputs so that you do not accidentally mark dangerous inputs as untainted (see CWE-183 and CWE-184).
-
Implementation
For Python programs, it is frequently encouraged to use the ast.literal_eval() function instead of eval, since it is intentionally designed to avoid executing code. However, an adversary could still cause excessive memory or stack consumption via deeply nested structures [REF-1372], so the python documentation discourages use of ast.literal_eval() on untrusted data [REF-1373].
Source: MITRE CWE-94
1,733 results · Clear filters · Feed
-
High 7.1
virtualenv: Xonsh activation literals depend on the script decoding codec
virtualenv is a tool for creating isolated virtual python environments. Prior to 21.14.9, POSIX xonsh activation quote() uses repr() to write printable non-ASCII prompt or destination data into UTF-8 activate.xsh, while xonsh can read that script using $XONSH…
-
High 8.6
A vulnerability could allow a remote authenticated attacker to achieve arbitrary code execution with privileges on the …
-
High 7.8
An issue in nvbn thefuck v.3.32 allows a local attacker to execute arbitrary code via the get_rules_import_paths in the…
-
High 7.2
Craft CMS 5.8.0 before 5.11.0 Admin RCE via ContentBlock Field Layout
Craft CMS from 5.8.0 before 5.11.0 contains a remote code execution vulnerability that allows authenticated administrators to inject Yii behaviors by supplying a JSON-string field layout to ContentBlock::setFieldLayout(). Attackers with allowAdminChanges enab…
-
High 8.8
Craft CMS 5.0.0 before 5.11.0 RCE via Asset::getMimeType Transform Injection
Craft CMS from 5.0.0 before 5.11.0 contains a remote code execution vulnerability that allows authenticated Control Panel users to escape the Twig sandbox via Asset::getMimeType transforms. Attackers can supply 'as' or 'on' keys passed unfiltered to Craft::cr…
-
High 8.8
Craft CMS 5.0.0 before 5.11.0 RCE via Twig Sandbox Behavior Injection
Craft CMS from 5.0.0 before 5.11.0 contains a remote code execution vulnerability that allows authenticated Control Panel users to escape the Twig sandbox by passing uncleansed criteria to getNext, getPrev, or getOwners. Attackers can inject Yii "as" or "on" …
-
High 8.8
Craft CMS before 4.18.8 and 5.11.0 Authenticated RCE via Twig group Filter Array-Callable
Craft CMS 4.x before 4.18.8 and 5.x before 5.11.0 contains a remote code execution vulnerability in the Twig group filter, which invokes non-string arrow arguments as callables without the checkArrow guard used by other arrow-accepting filters. Authenticated …
-
High 7.3
sqlparse through 0.6.0 PHP Code Injection via output_format='php'
sqlparse through 0.6.0 contains a PHP code injection vulnerability that allows attackers to inject PHP expressions by supplying SQL text containing unescaped dollar signs. Attackers can embed ${expr} or {${expr}} sequences that OutputPHPFilter places in doubl…
-
High 8.8CVE-2026-83531 Public exploit
Apache Camel: Camel-Dynamic-Router Control Endpoint Compiles Predicate Expressions Supplied by Control Messages
Improper Control of Generation of Code ('Code Injection'), Improper Neutralization of Special Elements used in an Expression Language Statement ('Expression Language Injection') vulnerability in Apache Camel. This issue affects Apache Camel: from 4.4.0 before…
-
High 8.1
User Registration & Membership <= 5.2.8 - Authenticated (Subscriber+) Arbitrary Shortcode Execution via Referer Header via 'referrer_url' Smart Tag
The User Registration & Membership plugin for WordPress is vulnerable to Arbitrary Shortcode Execution in versions up to, and including, 5.2.8 This is due to insufficient input sanitization in the 'referrer_url' branch of UR_Smart_Tags::process(), where the a…
-
High 8.5
TuxGuitar: Arbitrary Code Execution via maliciously crafted `.tg` file
TuxGuitar is an open-source guitar tablature editor. Prior to 2.1.0, an untrusted .tg file can set gervil.soundbank.path to reference an on-disk archive, including the .tg file itself, whose META-INF/services/javax.sound.midi.Soundbank entry names an attacker…
-
High 8.8
Apache Kylin: JDBC URL Whitelist Verification Bypass
Improper Control of Generation of Code ('Code Injection') vulnerability in Apache Kylin. Special parameter in JDBC URL can lead to code injection. Impact most to the kylin services that are directly exposed to public network, which is not common. In-house ky…
-
High 8.8
An issue in the context-compaction feature of DeepSeek Harness (dsh) commit c291e796 allows attackers to execute a prom…
-
High 8.5
Markdown Preview Enhanced: Remote Code Execution via VM Sandbox Escape in `.crossnote/config.js` and `.crossnote/parser.js`
Markdown Preview Enhanced is a Markdown preview extension for Visual Studio Code. Prior to 0.8.29, Markdown Preview Enhanced evaluates repository-controlled .crossnote/config.js and .crossnote/parser.js files with vm.runInNewContext, which does not provide a …
-
High 7.5
WordPress Salevo plugin <= 1.3.0 - Content Injection vulnerability
Improper Control of Generation of Code ('Code Injection') vulnerability in VillaTheme Salevo woo-boost-sales allows Code Injection.This issue affects Salevo: from n/a through 1.3.0.
-
High 8.5
PAN-OS: Authenticated Remote Code Execution (RCE) in the Management Interface
A remote code execution (RCE) vulnerability in the management interface of Palo Alto Networks PAN-OS® software enables an authenticated attacker with network access to execute arbitrary code with root privileges on the firewall. The security risk posed by thi…
KEV Oct 14, 2026 -
High 8.8
Gitea reusable workflow secret exposure in pull_request_target
Gitea resolves local reusable workflow files from a fork's head commit when a trusted base-branch workflow runs for a pull_request_target event. A fork author can modify a referenced reusable workflow so attacker-controlled code executes with the repository a…
-
High 7.5
WordPress Email Template Customizer for WooCommerce plugin <= 1.3.4 - Content Injection vulnerability
Improper Control of Generation of Code ('Code Injection') vulnerability in VillaTheme Email Template Customizer for WooCommerce email-template-customizer-for-woo allows Code Injection.This issue affects Email Template Customizer for WooCommerce: from n/a thro…
-
High 7.2
Affiliate Toolkit 3.8.7 - 3.9.0 - Contributor+ RCE via BladeOne @php Directive
The affiliate-toolkit WordPress plugin before 3.9.1 does not prevent a low-privileged user from supplying a template body that its bundled templating engine compiles and executes as PHP, allowing users with the Contributor role or above to run arbitrary PHP c…
-
High 8.8
TEAMPLUS TECHNOLOGY｜TeamPlus Messenger - Code Injection
TeamPlus Messenger developed by Teamplus Technology has a Code Injection vulnerability. Unauthenticated remote attackers can trick users into interacting with crafted content to inject and execute arbitrary JavaScript. After the user confirms a Windows securi…
-
High 8.8
TEAMPLUS TECHNOLOGY｜TeamPlus Messenger - Code Injection
TeamPlus Messenger developed by Teamplus Technology has a Code Injection vulnerability. Unauthenticated remote attackers can trick users into opening a crafted link to inject and execute arbitrary JavaScript. After the user confirms a Windows security warning…
-
High 8.8
ASP.NET Core Remote Code Execution Vulnerability
Improper control of generation of code ('code injection') in ASP.NET Core allows an unauthorized attacker to execute code over a network.
-
High 7.8
Windows Assigned Access Elevation of Privilege Vulnerability
Improper control of generation of code ('code injection') in Windows Assigned Access allows an authorized attacker to elevate privileges locally.
-
High 7.8
Windows Performance Monitor Elevation of Privilege Vulnerability
Improper control of generation of code ('code injection') in Windows Performance Monitor allows an authorized attacker to elevate privileges locally.
-
High 8.8
Windows Encrypting File System (EFS) Remote Code Execution Vulnerability
Improper control of generation of code ('code injection') in Windows Encrypting File System (EFS) allows an authorized attacker to execute code over a network.
-
High 8.8
Microsoft Office SharePoint Elevation of Privilege Vulnerability
Improper control of generation of code ('code injection') in Microsoft Office SharePoint allows an authorized attacker to elevate privileges over a network.
-
High 8.8
Visual Studio Tools for Applications and SQL Server Management Studio Remote Code Execution Vulnerability
Improper control of generation of code ('code injection') in Visual Studio Tools for Applications and SQL Server Management Studio allows an unauthorized attacker to execute code over a network.
-
High 8.8
dataclasses-avroschema enum symbols can inject statements through generated member names
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.38.0 through 0.70.5, BaseGenerator.parse_enum and EnumRepresentation.render_symbols use each …
-
High 8.8
dataclasses-avroschema allows code injection through fixed-type aliases supplied as a bare string
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.38.0 through 0.70.5, BaseGenerator.parse_fixed interpolates a fixed type's aliases value dire…
-
High 8.8
dataclasses-avroschema enum field defaults can inject statements into generated modules
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.38.0 through 0.70.5, FieldRepresentation.get_field_default renders a schema-controlled enum f…
-
High 8.8
dataclasses-avroschema allows code injection through generated enum-symbol string literals
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.38.0 through 0.70.5, EnumRepresentation.render_symbols concatenates each schema-controlled en…
-
High 8.8
dataclasses-avroschema pydantic-class metadata allows code injection in generated annotations
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.44.0 through 0.70.5, the Pydantic and AvroDantic model generators insert a schema field's pyd…
-
High 8.8
dataclasses-avroschema field documentation allows code injection in generated Pydantic descriptions
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.65.9 through 0.70.5, PydanticFieldRepresentation.add_field_properties inserts a schema field'…
-
High 8.8
dataclasses-avroschema fixed namespaces allow code injection in generated annotations
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.38.0 through 0.70.5, BaseGenerator.parse_fixed inserts a fixed type's namespace value between…
-
High 8.8
dataclasses-avroschema record documentation can terminate generated class docstrings
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.40.0 through 0.70.5, ClassRepresentation.render_docstring copies a schema record's doc value …
-
High 8.8
dataclasses-avroschema allows code injection through generated Meta.schema_name
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.62.3 through 0.70.5, ClassRepresentation.render_metaclass inserts a schema-controlled Avro re…
-
High 8.8
dataclasses-avroschema allows code injection through record names in generated default factories
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.47.3 through 0.70.5, FieldRepresentation.get_field_default inserts a schema-controlled refere…
-
High 8.8
dataclasses-avroschema allows code injection through generated field-default literals
dataclasses-avroschema generates Avro schemas from Python dataclasses, Pydantic models, and Faust Records and generates Python code from Avro schemas. From version 0.38.0 through 0.70.5, FieldRepresentation.get_field_default inserts schema-controlled string, …
-
High 8.8
heym Python Tool Sandbox Escape
Heym builds and runs agentic systems and business-process automations on self-hosted infrastructure. Prior to 0.0.40, an authenticated user who can create or edit a workflow with an Agent custom Python tool can bypass the blacklist-based sandbox in backend/ap…
-
High 8.8
DataStage on Cloud Pak for Data Code Injection
IBM DataStage on Cloud Pak for Data 5.4.0.0 could allow a remote authenticated attacker to execute arbitrary code due to improper neutralization of special elements used in an OS command.