Description

TypeBot is a chatbot builder tool. Version 3.16.1 has a CSV injection vulnerability in the result export functionality. The application does not sanitize or escape user-supplied input when generating CSV files. An attacker can inject spreadsheet formulas into input fields, which are later executed when an administrator opens the exported CSV in spreadsheet software such as Microsoft Excel or LibreOffice Calc. Version 3.17.0 patches the issue.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "9.6",
  "pubDate": "2026-08-11T18:17:27.123Z",
  "pubdate": "2026-08-11T18:17:27.123Z",
  "executiveSummary": "TypeBot version 3.16.1 suffers from a CSV injection vulnerability within its result export functionality.\nThe security flaw arises due to the application's failure to properly sanitize or escape user-supplied input strings prior to incorporating them into generated comma-separated values (CSV) export files.\nConsequently, malicious actors possessing the capability to submit input data to chatbot instances can inject arbitrary spreadsheet formulas, such as DDE (Dynamic Data Exchange) commands or command execution sequences.\nWhen an administrator or privileged user subsequently exports the response data and opens the resulting CSV file utilizing vulnerable spreadsheet software such as Microsoft Excel or LibreOffice Calc, the injected formulas are automatically evaluated and executed within the context of the user's operating system session.\nThis introduces significant risk implications, potentially leading to local privilege escalation, arbitrary command execution, or unauthorized data exfiltration from the administrator's local machine.\nExploitation requires the attacker to successfully inject malicious payloads into chatbot input fields and necessitates social engineering or administrative action to trigger the export and opening of the CSV file in a susceptible spreadsheet application.",
  "technicalDetails": "The root cause of the vulnerability resides in the lack of input sanitization and output encoding mechanisms within the result export module of TypeBot.\nWhen processing user-supplied data for CSV generation, the application directly serializes raw input fields that begin with dangerous control characters—such as equals signs (=), plus signs (+), minus signs (-), or at symbols (@)—without preceding them with defensive characters like single quotes or neutralizing the syntax.\nThe vulnerable component is specifically the export functionality handling user-submitted conversational inputs in TypeBot version 3.16.1.\nTo exploit this flaw, an unauthenticated or authenticated attacker interacts with a deployed TypeBot chatbot instance and supplies a specially crafted string containing a malicious spreadsheet formula as conversational input.\nThis payload is stored persistently within the underlying database without validation or sanitization.\nAt a later stage, an administrative user initiates a data export operation, instructing the application to generate a CSV file containing the stored conversation results.\nThe application compiles the raw, unsanitized user inputs into the standard CSV format and serves the file for download.\nWhen the administrator opens this CSV file using spreadsheet software like Microsoft Excel or LibreOffice Calc, the application parses the leading characters of the malicious field as a formula rather than static string data.\nThe spreadsheet processor then evaluates the injected syntax, which can trigger local application spawning, command execution via command interpreters (e.g., cmd.exe or bash), or requests to external servers containing sensitive environment data.\nThe post-exploitation impact depends heavily on the privileges of the user opening the file and the configuration of the local spreadsheet software, potentially enabling full system compromise of the administrator workstation."
}