> For the complete documentation index, see [llms.txt](https://docs.ox.security/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.ox.security/vibesec/malicious-dependency-governance.md). # Malicious Dependency Governance > **Note:** > > * This capability is currently in Early Access (EA) and is not generally available. To request access, please contact OX technical support. Dependency governance helps prevent AI coding agents from installing known malicious dependencies. VibeSec detects dependency installation attempts, validates the requested packages and versions, and blocks the operation when a malicious dependency is identified. You can review both blocked and validated dependency activity in the OX platform. ### How dependency validation works VibeSec validates dependencies when an AI agent attempts to add them by: * Running a package installation command, such as `npm install`. * Editing or writing to a dependency manifest file. When OX identifies a malicious package or version, it blocks the operation, and an error message appears in the AI coding tool. When a single installation includes multiple dependencies, OX evaluates each dependency individually. However, if one dependency is malicious and the dependencies are part of the same command or file update, OX blocks the entire operation. No dependencies from that operation are installed. {% hint style="info" %} Currently, dependency validation supports packages installed through the npm package manager (the npm ecosystem). {% endhint %} ### Review dependency activity **To review dependency installation attempts:** * Go to the **Agent Activity Log** and filter by **Dependency installation**.
The log includes both blocked and validated dependency activity. | Column | Description | | ------------------ | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Activity** | The type of activity. Dependency validation events appear as **Dependency installation**. | | **Details** | The dependency and version that the agent attempted to add. When the operation includes multiple dependencies, this column displays the number of dependencies. | | **Risk Mitigated** | The risk detected by OX. For malicious dependencies, this value is **Restricted dependency blocked**. | | **OX Action** |

The action taken by OX:

| | **When** | When the activity occurred. | | **User** | The user associated with the agent activity. | | **Tool** | The AI coding tool in which the activity occurred. | {% hint style="info" %} The Agent Activity Log displays activity from the last seven days. Older events are not retained. {% endhint %} ### Review dependency installation details Select a dependency installation event to review why OX allowed or blocked the operation and which dependencies were involved.
The details page includes the following sections. | Section | Description | | --------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Event summary** | Displays the dependency installation summary, including the user, user role, AI coding tool, event time, and the OX action that was taken. | | **Blocked by Policy** | Explains why the operation was blocked. If multiple dependencies were included in the same installation and one violated policy, OX blocks the entire installation and no dependencies are installed. This section is displayed only for blocked events. | | **Dependencies** |

Lists each dependency that the agent attempted to install, including the package name, requested version, package ecosystem, and validation status.

The following statuses are available:
- Malicious: OX identified the dependency or requested version as malicious.
- Cleared: OX validated the dependency and did not identify it as malicious.

| | **Event Metadata** | Displays technical information about the event, including the tool that attempted the installation and the event ID. | {% hint style="info" %} If a single installation command includes both malicious and non-malicious dependencies, OX blocks the entire operation. No dependencies from that installation are added. {% endhint %} ### Review blocked dependencies in Risks Prevented page The Risks Prevented page provides a focused view of agent activity for which OX blocked a risk. Dependency installation events appear on this page only when the dependency was blocked. Validated dependency installations are available in the Agent Activity Log, but do not appear in Risks Prevented.
{% hint style="info" %} Risks Prevented displays activity from the last seven days. Older events are not retained. {% endhint %} ### Enable dependency governance 1. Go to **Settings > VibeSec**.
2. Under **AI Governance**, make sure **Dependencies** is ON (Default: ON). --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.ox.security/vibesec/malicious-dependency-governance.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.