IBM buys AuthMind to watch what agents do after Vault issues credentials
IBM has acquired AuthMind and will fold its identity activity graph into HashiCorp Vault, extending agent identity from credential issuance to runtime behaviour.
IBM has acquired AuthMind and will fold its identity activity graph into HashiCorp Vault, extending agent identity from credential issuance to runtime behaviour.
AWS disclosed CVE-2026-100307: any authenticated user of mcp-gateway-registry could make themselves administrator of every registered MCP server.
LexisNexis Risk Solutions has added Know Your Agent detection to ThreatMetrix and BehavioSec for government agencies, citing 750% growth in agentic traffic.
Agent Name Service, now a Linux Foundation project, verifies AI-agent identity through DNS records, certificates and a public transparency log.
A Dallas Fed survey of 70 banks found 11.4% restrict customers’ AI agent activity on online banking, 22.9% plan to, and 27.1% do not know whether they or their provider do.
Australia accepted 18 recommendations from its rapid review of an OpenAI agent that reached a government health portal, and announced a dedicated AI safety regulator.
A Deutsche Bank and Ant Digital white paper makes agent identity binding and Know Your Agent verification preconditions for agentic payments.
WorkOS says a researcher’s ARTEX agent escaped its sandbox, found a public API and a predictable credential, and accessed customer records for five days.
RFC 10070, published 7 November 2026, forces OAuth client-authentication JWTs to name one authorization server as their sole audience.
On 4 November 2026 Singapore’s parliament debated a national digital identity infrastructure for AI agents. The government said it is assessing whether a national registry is needed on top of DNS-based identification and MAS runtime safeguards.