What Happened
Varonis describes ForcedLeak as a critical vulnerability chain in Salesforce Agentforce that enables silent CRM data theft via indirect prompt injection, agent overreach, and content security policy misconfiguration.[10] The exploit allows attackers to use crafted prompts to manipulate the agent into exfiltrating sensitive customer relationship data while bypassing standard access controls.[10]
Why It Matters
Report facts: Varonis describes ForcedLeak as a vulnerability chain in Salesforce Agentforce that uses indirect prompt injection, agent overreach, and content security policy misconfiguration to silently exfiltrate sensitive CRM data while bypassing normal access controls. Attackers craft prompts that manipulate the agent into leaking customer relationship data without obvious signs to the victim organization. RealGround analysis: This reflects a high-severity indirect prompt injection and AI agent abuse scenario where business logic and access control are effectively bypassed through content-level attacks. Organizations should harden agent decision logic, strictly constrain data-access actions, and continuously red team AI agents against malicious prompt patterns that could trigger unauthorized data leakage.
RealGround Analysis
This signal maps to indirect prompt injection. Organizations using AI agents, LLM APIs, SaaS integrations, or sensitive data workflows should review whether this class of issue could create unauthorized tool execution, data leakage, weak approval gates, or unmanaged supply-chain exposure.
Recommended Actions
- Restrict AI agent tool permissions and production write paths.
- Review sensitive data access across prompts, logs, embeddings, memory, and SaaS integrations.
- Add human approval workflows for high-impact or state-changing actions.
- Run prompt injection and indirect prompt injection tests against affected workflows.
- Document the owner, control gap, and remediation deadline for this risk class.