--- title: "Automatic classification with remediation scans" slug: "automatic-classification-with-remediation-scans-3" tags: ["Product Usage", "SaaS"] updated: 2026-09-01T09:28:07Z published: 2026-09-01T09:28:07Z canonical: "docs.trellix.com/automatic-classification-with-remediation-scans-3" --- > ## Documentation Index > Fetch the complete documentation index at: https://docs.trellix.com/llms.txt > Use this file to discover all available pages before exploring further. # Automatic classification with remediation scans Remediation scans can classify files by embedding a classification in the file's properties. You can classify files as part of a remediation action. The classification is embedded into the file's properties. As a result, the classification is not lost when the files are changed, moved between repositories, uploaded to the web, or sent by email. To classify files with remediation scans, configure the scan using a network discovery rule with a **Classify File As** action. When the repository is scanned, files that meet the conditions configured in the rule are classified according to the selected classification. Also, the classification's **Tag ID** is embedded into the file's properties. To be classified, files must have one of these conditions specified in the rule: keyword, dictionary, advanced pattern, proximity, or document properties. The classification applied by the rule is selected on the **Reaction** tab, allowing the user to set the **Classify File As** classification to be: - The same as the classification on the **Condition** tab. - A different classification from the **Condition** tab classification with criteria. - A different classification from the **Condition** tab classification without any criteria (for labeling). ### Limitations The following limitations apply to applying classifications with remediation scans: - A rule can have only one classification configured in the **Classify File As** reaction. - A file is limited to five embedded classifications. (That is, up to five rules can apply to one file.) - The user credentials configured in the remediation scan must have read, write, and modify permissions. - Protected or secured files (when **Modify** isn't allowed) can't be classified. - Only the following file extensions are supported: | aif | mpeg | tiff | | --- | --- | --- | | aiff | mpg | vsd | | avi | msg | vsdx | | dng | pdf | vss | | doc | png | vst | | docm | pot | wav | | docx | potm | wma | | dot | potx | wmv | | dotm | pps | wps | | dotx | ppsm | xdcam | | eps | ppsx | xls | | jpeg | ppt | xlsb | | jpg | pptm | xlsm | | mov | pptx | xlsx | | mp2 | ps | xlt | | mp3 | psd | xltm | | mp4 | swf | xltx | | mpa | tif | xps | > [!NOTE] > Note > > For unsupported file types, the operational event `File type is not supported for auto classification.` is sent. - The content of the file must match its extension or it can't be classified. For example, a .txt file with extension changed to .doc is not classified. - PostScript files (*.ps and *.eps) must be at least version 3.0. The standard library doesn't support versions earlier than 3.0, so they are not classified. - MP3 files contain metadata inside a special ID3 tag. This tag usually contains information like track title, artist, and album. For the files to be classified, this tag must exist. MP3 files without such tag are not classified. - *.mpeg and *.mpg files must be of ISO media file format to be classified. To verify this, open the file with Notepad. After the first 4 bits is some text starting with `ftyp`, for example `ftypmp42`, if the file format is ISO. - Files with size=0 are not downloaded, so can't be tagged. The operational event `File is not supported.` is sent. ### Removing automatic classifications Remediation scans in File Server repositories can remove automatic classifications and apply them. Creating a rule with a **Remove Automatic Classification** action removes the specified classification tag from a file without affecting other tags that might be applied to the file. The rule can specify only one classification, and can't remove tags applied manually.