Question 730
At Meridian Research, a research support office is improving laboratory service intake during a controlled sandbox pilot (SF-M05-02). The administrator observes that an Agentforce agent is granted a generic MCP tool endpoint that can mutate unrelated customer systems. Any fix must ensure that restricted project data remains separated by research group. Which action best addresses the issue while preserving the stated requirement?
Explanation
Correct Answer
The documented failure is that an Agentforce agent is granted a generic MCP tool endpoint that can mutate unrelated customer systems. Restrict the MCP tool access to approved operations with identity and authorization checks for Meridian Research's laboratory service intake workflow (SF-M05-02) while respecting that restricted project data remains separated by research group. Its outcome is not established merely by changing instructions: In a controlled Lab_Request__c acceptance test, attempts to invoke forbidden tools fail and allowed calls produce audit records; the record is retained under Meridian Research's rollout (SF-M05-02). The underlying reason is that Open tool protocols do not remove the need for least-privilege access boundaries. In this laboratory service intake scenario, this separates supported agent behavior from assumptions about the AI model.
Incorrect Answers
The action “Require authenticated agent identities and enforce recipient-side authorization for delegated tasks; implement it with a controlled rollback path” addresses a separate source case: A specialist agent trusts a delegated instruction to reveal private data outside its permitted audience. The correction must ensure that citizen information stays limited to the relevant case. Its underlying reason is: Cross-agent messages are requests, not blanket grants of authority. This question instead describes an Agentforce agent is granted a generic MCP tool endpoint that can mutate unrelated customer systems. The correction must ensure that restricted project data remains separated by research group. The original answer “Restrict the MCP tool access to approved operations with identity and authorization checks; implement it and keep the business rule enforceable” targets that failure because open tool protocols do not remove the need for least-privilege access boundaries.
The action “Use A2A for agent-to-agent collaboration and MCP where a client needs access to external tools or resources; implement it with least-privilege execution” addresses a separate source case: An architect proposes a tool-connection protocol for communicating between independent conversational agents. The correction must ensure that guest profile fields are used only with permitted access. Its underlying reason is: A2A and MCP serve distinct interoperability relationships. This question instead describes an Agentforce agent is granted a generic MCP tool endpoint that can mutate unrelated customer systems. The correction must ensure that restricted project data remains separated by research group. The original answer “Restrict the MCP tool access to approved operations with identity and authorization checks; implement it and keep the business rule enforceable” targets that failure because open tool protocols do not remove the need for least-privilege access boundaries.
The action “Keep a single appropriately scoped agent when multi-agent orchestration adds no control or scale benefit; implement it before the next environment promotion” addresses a separate source case: A basic FAQ workflow has one data source and no specialized actions yet architects propose six coordinating agents. The correction must ensure that supplier prices are never shown outside entitled teams. Its underlying reason is: Multiple agents should be chosen for a business or control need, not novelty. This question instead describes an Agentforce agent is granted a generic MCP tool endpoint that can mutate unrelated customer systems. The correction must ensure that restricted project data remains separated by research group. The original answer “Restrict the MCP tool access to approved operations with identity and authorization checks; implement it and keep the business rule enforceable” targets that failure because open tool protocols do not remove the need for least-privilege access boundaries.