Incident report · Prompt injection

Word Worm Crawls into Copilot, Spreads Chaos

Håkon Måløy hid instructions in a Word document as small white text and checked whether Copilot would carry them into the next document it generated. It did: the payload altered figures in the resulting reports and copied itself onward without the victim noticing. Måløy says two mitigation attempts, including a model upgrade, did not close the vulnerability class; Microsoft says it has addressed the findings.

Discovered
Jul 29 2026
Failure mode
Prompt injection
Impact
Financial figures altered in generated reports · Self-propagating injection across Microsoft 365 documents

Summary from the public record. The failure mode’s analysis follows.

The agent did exactly what it was told. The problem was who was doing the telling.

A failure mode to test

The agent could not tell the difference between text it was reading and instructions it should follow. Once untrusted content reached its context window, it carried the same weight as the operator's own prompt, and the agent acted on it with every permission it had.

This incident illustrates a failure mode to test. Whether a control stops it depends on the integration, policy and execution path; no Fencio prevention test is claimed here.

Say yes to the AI your business wants.

Book a 30-minute call. Bring your agents and your rules.