braintrust-write-agent-behavior-spec

Warn

Audited by Snyk on Aug 20, 2026

Risk Level: MEDIUM
Full Analysis

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

  • Third-party content exposure detected (medium risk: 0.30). [The required workflow reads and classifies user/attacker-provided free-text contained in “Traces” (LLM judge reads text the system under test produced), which is outsider-authored when the attacker can influence the system’s outputs; this creates an indirect prompt-injection exposure via trace content. The documents explicitly warn that trace content includes user turns and tool outputs shaped like instructions.]

Issues (1)

W011
MEDIUM

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

Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 20, 2026, 10:16 PM
Issues
1