NYC

agentforce-2025

Warn

Audited by Snyk on Feb 15, 2026

Risk Level: MEDIUM
Full Analysis

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.90). The skill explicitly ingests and processes untrusted user-generated content (Web Chat, SMS/WhatsApp, Email, Slack) and also calls external LLM APIs (OpenAI/Anthropic callouts shown in "OpenAI GPT Integration" and "Anthropic Claude Integration"), which the agent reads and interprets as part of its workflow—exposing it to indirect prompt injection.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

  • Potentially malicious external URL detected (high risk: 0.80). This skill issues runtime HTTP callouts to external LLM endpoints (callout:OpenAI/v1/chat/completions and callout:Anthropic/v1/messages), and the model outputs returned by those endpoints are used directly as agent responses/instructions, so they are runtime dependencies that can control prompts.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 15, 2026, 08:57 PM