IBM Launches Agent Identity and Control Stack for AI Agents with HashiCorp Vault
Announced at TechXchange, the stack adds per-agent identity records, short-lived least-privilege credentials, sandboxing and a kill switch.
IBM announced a stack for AI agent identity and control at TechXchange on 27 October. Agent Identity, built with HashiCorp Vault, creates an authoritative record per agent covering owner, lifecycle and authorities, paired with Vault-based credential controls: least-privilege access, short-lived credentials and real-time revocation when an agent attempts an unauthorised action. The watsonx Orchestrate additions include Discover, a tech preview that finds agents built on third-party platforms such as Amazon Bedrock, Microsoft Foundry and Google Gemini so they can be managed; Contain, a tech preview providing an isolated sandbox per agent with a kill switch; and Evaluate, which tests agent, model and MCP server quality and compliance before and after deployment with per-agent cost tracking.
IBM frames the release with Gartner-derived context that the average Fortune 500 will use more than 150,000 agents by 2028, while only 13% of organisations consider their agent governance adequate. Integration with NVIDIA OpenShell combines protected runtimes with verified identity and short-lived credentials per action. Availability and pricing were not detailed in the announcement.