Visibility into AI application sharing
Claroc describes its service as monitoring SaaS and AI application adoption, sharing and configuration before and after approval. Its GhostShares report examined a concrete exposure mechanism: Meta Muse placed share identifiers and conversation titles in hostnames discoverable through passive-DNS and subdomain tools. The researchers reported opening a sample without logging in. This was an investigation of publicly shared links and their discoverability, not evidence that every private Muse conversation was exposed. [1]
The report was published October 18, 2026 and updated October 20. In that update, Claroc said none of 40 previously accessible links opened during a recheck, while historical hostnames and titles remained in passive-DNS records. Ars Technica's October 20 coverage carried Meta's statement that it had paused sharing while reviewing the report. That documents a response at the time, without establishing all later sharing behavior. [1][2]
Sources
- GhostShares: A Shared Muse Chat Exposed Prep for a Confidential Call With a Frontier AI Lab Leader
Claroc · Oct 18, 2026
- Meta Muse exposes users’ personal and business secrets when they share conversations
Ars Technica · Oct 20, 2026