At Dreamforce, the company introduced an interface layer that can be accessed via API, MCP or command line, and security tools that scan MCP servers at the time of registration.
Presented at Dreamforce, held from September 15 to 17 in San Francisco, AIforce is described by Salesforce as a composable interface layer that carries the rest of the platform on any surface. The idea is to allow employees and software agents to work on the same data and under the same governance, whether access is through an API, the MCP protocol or a command-line interface — so without necessarily opening the Salesforce application itself. The underlying technical layer, Headless 360, was announced on August 19.
Two bricks accompany the announcement on the construction side: Builder Central, in beta, brings together in a single point the implementation, development and deployment of solutions, with automated security controls and the possibility of starting from a natural language instruction; the Headless Experience Layer, now generally available, automatically translates business rules configured only once into native and secure interfaces on each surface.
Agent safety goes into the product
Two functions explicitly target agent-specific risks. Security Mesh provides a standardized view for spotting anomalous API activity and compromised accounts. MCP Risk Scores, on the other hand, automatically scan MCP servers as they are registered in Agentforce, looking for prompt injections and tool poisoning attacks—an acknowledgement in a commercial catalog that the agent interconnect protocol is now an attack surface in its own right.
Read the full article on the original source
Go to the source →