privileged-access
Warn
Audited by Snyk on Jul 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: 0.75). The required workflow is an assessment that explicitly treats “vault metadata, session logs, and policy descriptions” as untrusted input (Step 1–6), which are outsider-authored free text originating from third-party PAM systems and logs that the agent would ingest into the LLM context.
Issues (1)
W011
MEDIUMThird-party content exposure detected (indirect prompt injection risk).
Audit Metadata