Executive Summary
Three 'SalesBleed' vulnerabilities in Salesforce Agentforce let attackers hijack trusted agents, steal data and launch phishing attacks without user clicks.
What Happened
Three vulnerabilities in Salesforce Agentforce allowed hackers to hijack trusted agents, steal data, and launch phishing attacks. The post ‘SalesBleed’ Flaws in Salesforce Agentforce Enabled Zero-Click Data Exfiltration appeared first on SecurityWeek .
Affected Products
- Salesforce Agentforce
Vulnerability / Attack Technique
Zero-click agent hijacking via injected instructions
Technical Mechanism
Attacker-controlled data reaching the agent caused it to act on the attacker's instructions, exfiltrating data and sending phishing.
AI Security Category
LLM01:2025 Prompt Injection
Part of the Prompt Injection topic.
Relevant Frameworks
RealGround Analysis
A trusted business agent became an attacker's tool for data theft and phishing. RealGround analysis: agents that can both read untrusted data and send messages need output controls.
Recommended Controls
- Apply Salesforce's fixes
- Limit Agentforce's ability to send external messages
- Review which data sources feed agent context
Related RealGround Service
AI Agent Business Logic Audit · or start with the free AI Agent Security Scanner.
Sources
securityweek.com: https://www.securityweek.com/salesbleed-flaws-in-salesforce-agentforce-enabled-zero-click-data-exfiltration/
Summary and analysis written by RealGround from the source above; last updated 2026-10-17.