Skip to content
Get Ready for Agentsby Prefactor
ProductsAgent SignalsAgent Ready
NewsAgent NewsAgent Failures
Podcast
Live IndicesAgent IndexAgentic AI Jobs IndexJobs Index — Data & MethodologyAI Governance IndexAI Conference Index
ExploreAgentic AIAI AgentsAI GovernanceAI SecurityCrewAIGenerative AILangChainLangGraphMCPPrompt EngineeringRAGCompareGuidesMaturity ModelAI Agent Stats
Agentic AIAgentic AI overviewAgentic AI vs generative AIAgentic AI vs AI agentsAgentic AI vs traditional automationAgentic AI architecture: the loop and its partsAgentic AI frameworks: the landscape and how to chooseAgentic AI examples: where the loop does real workAutonomous AI agents: what autonomy actually meansMulti-agent systems: when one agent becomes several
AI AgentsAI Agents overviewAgentic AI vs AI agentsAI agents vs chatbotsAI agents use casesWhat is an AI agent workflow?How do AI agents work?What is an AI agents framework?AI agents for automationAI agents for customer service
AI GovernanceAI Governance overviewRuntime governance vs pre-deployment reviewAI governance best practices that operate, not decorateWhat is AI governance?What is an AI governance framework?What is an AI governance policy?AI governance principlesAI ethics and governanceWhat is AI agent governance?
AI SecurityAI Security overviewAI security risks: a working taxonomyAI in cybersecurity: what it does, and what it doesn'tWill AI replace cybersecurity jobs?AI securityAI security best practicesEnterprise AI securityAI-Powered Cybersecurity Threats
CrewAICrewAI overviewCrewAI vs LangGraphCrewAI agentsCrewAI frameworkCrewAI memoryCrewAI use cases
Generative AIGenerative AI overviewAgentic AI vs generative AIGenerative AI modelsGenerative AI use casesWhat is generative AI?How does generative AI work?Generative AI Video Models
LangChainLangChain overviewLangChain vs LangGraphLangChain agentsLangChain memoryWhat is LangChain?LangChain tools
LangGraphLangGraph overviewLangChain vs LangGraphCrewAI vs LangGraphLangGraph agentsLangGraph architectureLangGraph deploymentLangGraph SDK
MCPMCP overviewMCP vs APIsHow to build an MCP serverHow to run MCP servers: local, remote, and hostedHow to secure MCP servers and clientsWhat is an MCP server?MCP architecture, explained
Prompt EngineeringPrompt Engineering overviewContext engineering vs prompt engineeringHow to do prompt engineeringPrompt engineering tutorialPrompt engineering techniques that survive contact with productionAdvanced prompt engineeringPrompt engineering for generative AIIs prompt engineering still relevant?AI prompt engineering
RAGRAG overviewRAG vs fine-tuningRetrieval-augmented generation tutorialRAG architecture: the pipeline and its decision pointsAgentic RAG: retrieval inside the loopRAG frameworkRAG use cases
Head to HeadLangChain vs LangGraphRBAC vs ABAC for AI agentsAgentic AI vs AI agentsRAG vs fine-tuningContext engineering vs prompt engineeringCrewAI vs LangGraphAll comparisons →
Hands-on GuidesHow to govern agentic AIPrompt engineering tutorialHow to implement agent observabilityHow to adopt an AI security framework that actually changes anythingHow to build an MCP serverHow to evaluate AI agentsAll guides →
Free ToolsAgent ReadyAgent Risk ConsultationAgent Maturity AssessmentMCP Maturity Audit
Prefactor LibraryMCP WhitepaperGlossaryBlog
CompanyAbout UsCareersContact UsNews & EventsPartnershipsSlack Community
Follow on Product HuntTake Assessment
Home/Agent Failures Index
Vendor toolingHighSecurity exploitData exposure18 Oct 2026

Muse's chat-sharing links were discoverable via passive DNS, exposing 6,800 conversations including loan applications and a bank fraud alert

Meta Muse users·Meta Muse

Muse built shared-conversation URLs by embedding the chat's own title into a subdomain of cf.metaaiusercontent.com; because that hostname gets logged in public passive-DNS records, security firm Claroc enumerated roughly 6,800 share links with a single passive-DNS query and opened hundreds of conversations users never intended to be public.

Blast radius
6,800 share links enumerable; a sample of opened conversations included a small-business loan application, a bank fraud alert with a partial account number, a company revenue review for a CEO, confidential executive-search notes, and a full-year staff vacation calendar. Claroc said it found no evidence of misuse.
Root cause
Using a user/AI-generated conversation title as part of a public identifier (subdomain) made nominally 'anyone with the link' shares discoverable without the link, instead of relying solely on an unguessable random token.
How it was caught
Claroc researchers found the links via a single passive-DNS search and reported it through Meta's bug bounty program on October 8.
What changed
Meta paused the Muse chat-sharing feature while it reviews the report and considers changes to the URL scheme.

Take-awayNever embed user or AI-generated content (like a title) into a public identifier such as a URL/subdomain for anything meant to stay private — DNS requests are effectively public record.

First surfaced: Security advisory · Recorded 21 Oct 2026
EngadgetYahoo TechSlashdot

Related incidents

  • Meta Muse appeared to reference a columnist's private Mac Messages despite his never granting accessJason Aten (Inc. Magazine columnist) · 30 Sept 2026
  • OpenAI's review finds agents leaked 53 user images and reached US government sitesOpenAI · 25 Sept 2026
  • Developers got Meta Muse to zip up and hand over its backing Linux filesystem, including internal Meta documentation, with minimal promptingMeta Muse users · 25 Sept 2026
  • MCP Atlassian server lets tool callers exfiltrate arbitrary server-local files via unrestricted attachment upload pathsUndisclosed · 10 July 2026

All 90 incidents in the Agent Failures Index →

Get Ready for Agentsby Prefactor

The hub for getting ready for AI agents — readiness scoring, the agentic jobs index, the Agent Reality Index, and the weekly Agent Brief, for the leaders driving agent adoption.

Assess

Readiness AssessmentMaturity ModelAgent Ready

Intel

BlogThe Agent BriefAgentic AI JobsJobs Index Data & MethodologyAgent IndexCompanies Building AI AgentsAgent Failures IndexGovernance IndexAI Conference Index

From Prefactor

Prefactor PlatformAgent Risk AuditMCP WhitepaperBlogGlossaryPodcast

Company

About UsCareersContact UsNews & Events
Privacy PolicyTerms of ServiceCookie PolicyAcceptable Use
© 2026 Prefactor Pty Ltd · Get Ready for Agents is a Prefactor project
Subscribe to the Agent Brief →Slack Community
Backed by Antler, Black Nova VC & Func Ventures