AGENTFORCE-21.2: Insecure Output Handling (Agent-to-XSS)
🔴 High · Data Exfiltration / Injection
Detects Flow screens targeted by agent actions that render LLM-generated output without sanitization. An adversarial LLM response containing HTML/JavaScript will execute in the Salesforce user's browser context. Tier B rule 34.1 extends this detection to full LWC component graph traversal.
Details
| Field | Value |
|---|---|
| Rule ID | AGENTFORCE-21.2 |
| Severity | High |
| Category | Data Exfiltration / Injection |
| Compliance | HIPAA, PCI_DSS, EU_AI_ACT_HIGH_RISK |
Remediation
Refer to the SquireX documentation for
remediation guidance specific to AGENTFORCE-21.2.