Microsoft Word Copilot Flaw Lets Hidden Prompts Spread Self

Gbhackers •Divya • July 30, 2026

Security researcher Håkon Måløy has disclosed a cross-domain prompt injection vulnerability affecting Microsoft Copilot for Word. This vulnerability could allow malicious instructions hidden within documents to alter Copilot-generated content and spread into newly created files. The issue, termed “Context Collapse, Part 3 – AI Worming through Word,” demonstrates how a document can serve as a […]

The full story

This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.

Self-Propagating AI Worm Discovered in Microsoft Word Copilot A critical vulnerability in Microsoft Copilot for Word allows hidden malicious prompts in documents to alter generated content and spread to new files. Disclosed by researcher Håkon Måløy, this issue, termed 'Context Collapse, Part 3 – AI Worming through Word,' has persisted despite Microsoft's mitigation attempts since March 2026. The vulnerability enables attackers to exploit untrusted documents without requiring access to a victim's Microsoft 365 tenant. Måløy's research indicates that the worm can propagate through normal workflows, making it difficult to trace its origin. Microsoft has acknowledged the issue but has not yet provided a robust solution. The attack vector relies solely on the inclusion of a compromised document in Copilot's context. The situation remains urgent as the worm continues to pose a significant risk to businesses using Microsoft Word. 6 sources 71

Extracted Entities

Microsoft Word Copilot Flaw Lets Hidden Prompts Spread Self

The full story

Extracted Entities

Companies (1)

Platforms (1)

Vulnerabilities (1)