Loading story…
A high-severity remote code execution flaw tracked as CVE-2026-73032 was disclosed in PapersGPT for Zotero affecting versions up to and including 0.6.1. The bug stems from passing untrusted LLM-generated content to window.eval() in views.ts, allowing arbitrary JavaScript execution inside Zotero's chrome-privileged context. Reported attack paths include prompt injection embedded in PDFs, man-in-the-middle tampering with API responses, and malicious custom LLM endpoints, with potential impact including file read/write access, process execution, and exposure of all Zotero data.
The maintainers addressed the issue by merging pull request #155, which removed window.eval from the execTag functionality and replaced another eval-based regex path with safer handling using new RegExp(...). The fix was merged into the main branch and tagged in papersgpt-v1.1.0, while public advisories classified the weakness as CWE-94 and noted proof-of-concept exploitation with high confidentiality, integrity, and availability impact.
Mallory correlates global threat intelligence with your attack surface — know if you’re exposed before adversaries strike.
5 events from the most recent confirmed update back to the earliest known activity.
An SSVC assessment was added to CVE-2026-73032 listing CISA Coordinator as the role and marking exploitation as proof-of-concept, not automatable, with total technical impact.
The CVE record for CVE-2026-73032 was newly received by disclosure@vulncheck.com and later modified the same day. The record identified PapersGPT for Zotero versions up to and including 0.6.1 as affected and added references to the repository, fixing commit, issue, pull request, and VulnCheck advisory.
VulnCheck published an advisory describing a remote code execution vulnerability in PapersGPT for Zotero 0.6.1 caused by unsanitized LLM output being passed to window.eval() in views.ts. The advisory documented exploitation vectors including prompt injection in PDFs, man-in-the-middle interception, and malicious custom LLM endpoints.
A verified commit removing window.eval of LLM responses in src/modules/views.ts was authored and merged via pull request #155 into the main branch. The patch also replaced eval-based regex handling and was associated with release tag papersgpt-v1.1.0.
A pull request was published to remove use of window.eval on LLM-generated output in the execTag function and replace another eval-based regex path with new RegExp, addressing a CWE-94 code injection issue in PapersGPT for Zotero.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
5 references tracked. Mallory keeps watching after this page renders.
cvefeed.io
Open sourcevulncheck.com
Open sourcegithub.com
Open sourcegithub.com
Open sourcegithub.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
A publicly disclosed flaw, **CVE-2026-46529**, allows single-click remote code execution in Evince-derived document viewers including **Atril** and **Xreader** by abusing PDF `/GoToR` actions. The bug stems from attacker-controlled PDF destination and search fields being interpolated into `ev_spawn()` and later reparsed by `g_app_info_create_from_commandline`, letting a malicious document inject a standalone `--gtk-module` argument. GTK then loads an attacker-controlled shared library, executing its constructor as the local user who opened the file. Advisories and mailing-list disclosures said the attack can be delivered as a single polyglot file that is both a valid PDF and a valid ELF shared object, making exploitation possible on stock Linux desktop installations without special configuration. Researchers reported proof-of-concept exploitation on Atril, including reverse-shell execution, and noted the issue is a variant of the defect class previously seen in `CVE-2023-51698`. Maintainers coordinated fixes across **Evince**, **Atril**, and **Papers**, while noting that GTK4-based viewers such as Papers are not vulnerable to the exact `--gtk-module` loading path because GTK 4 removed that flag, though related command-injection concerns may still remain.
Jun 29, 2026
Two newly cataloged CVEs detail local code execution vulnerabilities in Windows desktop applications **PDF Explorer 1.5.66.2** and **RGui 3.5.0**. **CVE-2018-25217** affects PDF Explorer through a structured exception handler (SEH) overflow in the application's **Custom fields settings** dialog, where malicious data placed in the **Label** field can overwrite SEH records and enable arbitrary code execution. The issue is mapped to `CWE-787` and was published with CVSS v4.0 and v3.1 scoring, alongside references to Exploit-DB, RTT Software, a trial installer, and a VulnCheck advisory. **CVE-2018-25258** affects RGui 3.5.0 through a stack-based buffer overflow in the **GUI preferences** dialog, specifically the **Language for menus and messages** field. The disclosure says an attacker can use SEH-based exploitation to bypass DEP, run a ROP chain that calls `VirtualAlloc`, and achieve arbitrary code execution on the local system. The entry was published with CVSS v4.0 and v3.1 vectors and references to the affected R 3.5.0 Windows binary, Exploit-DB, the R Project website, and a VulnCheck advisory.
Jun 29, 2026
A critical vulnerability (CVE-2025-65108, CVSS 10.0) has been identified in the `md-to-pdf` tool, which is used to convert Markdown files to PDF using Node.js and headless Chrome. The flaw allows attackers to achieve remote code execution by injecting malicious JavaScript into the front-matter section of Markdown files, exploiting the way the gray-matter library parses these blocks. This vulnerability is remotely exploitable and poses a significant risk to any system processing untrusted Markdown files with affected versions of `md-to-pdf`. The issue has been addressed in version 5.2.5 of the `md-to-pdf` library, and users are strongly advised to update to this version to mitigate the risk. No specific affected product versions are listed, but the vulnerability impacts all prior versions that use the vulnerable parsing mechanism. Security advisories have been published to highlight the severity and exploitation potential of this flaw, emphasizing the need for immediate remediation in environments where Markdown-to-PDF conversion is automated or exposed to user-supplied content.
Jun 29, 2026