Obsidian Security for Salesforce Agentforce is a security solution designed to protect AI agent workflows deployed within Salesforce Agentforce environments. It addresses security risks that arise when AI agents operate with excessive permissions, lack central governance, or interact with connected third-party applications. The solution provides visibility into Agentforce deployments by detecting and inventorying AI agents — including those created outside central governance (shadow agents). It monitors for permission misconfigurations, such as agents inheriting admin-level rights that bypass the Einstein Trust Layer, and provides actionable remediation guidance. Obsidian continuously monitors Agentforce workflows to identify data exposure risks stemming from AI-driven automation across use cases such as customer service, sales development, employee support, HR/IT help-desk, and operations scheduling. Key security concerns addressed include: - Excessive permissions: Agents inheriting elevated rights that create unnecessary data exposure - Shadow agents: Untracked agents created without central governance or ownership visibility - Supply chain risks: Connected app integrations that can introduce third-party attack vectors The platform aims to allow organizations to scale Salesforce Agentforce automation while maintaining security posture over AI-driven workflows and the SaaS applications they interact with.