Preliminary research

PromptFiction: a one-click flaw that made Claude Desktop act without consent

PromptFiction: Hijacking Claude Desktop Through the claude:// URL Scheme (Technical Report)

Elad Luz · pages.oasis.security

AI-collected research leads through 22 September 2026, including targeted additions between broader sweeps. Unranked, incomplete, not community-vetted, and subject to change.

Analyzes automatic instruction submission through Claude Desktop deep links. The report separates conversation-data access from filesystem effects requiring configured MCP tools and further permissions, and describes the revised user-confirmation boundary.

Record

Document

PromptFiction: Hijacking Claude Desktop Through the claude:// URL Scheme (Technical Report)

Researcher

Elad Luz

Published by

pages.oasis.security

Format

Whitepaper

Topic

Other

In the archive

• Collected in the 2026 preliminary research sweep

Related sources

• Claude Desktop Vulnerability: One-Click Prompt Injection

Tags

• prompt-injection
• ai-agent
• uri-scheme
• mcp
• owasp-a03-2021

This page is the archive's own catalogue record. The research is the work of Elad Luz, first published at the original source. Preserved copies are kept so the citation survives its host; this one was last captured on 18 September 2026.