Skills
Navigate
skills/jeremylongshore/tons-of-skills-marketplace/langfuse-enterprise-rbac/Gen Agent Trust Hub

langfuse-enterprise-rbac

Fail

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: CRITICALEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill references the official langfuse/langfuse image for deployment in the Docker Compose configuration instructions.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines procedures for processing external SAML assertions and organization configuration.
  • Ingestion points: SAML assertions handled in handleSAMLAssertion within references/implementation.md and environment variables defined in SKILL.md.
  • Boundary markers: The instructions include an explicit safety directive for the agent to "never include client secrets, API keys, or IdP assertions" in its generated access-control records.
  • Capability inventory: The skill operates within the scope of the Read, Write, and Edit tools.
  • Sanitization: The implementation code includes logic to validate that user emails in assertions match the expected organization domain.
  • [DATA_EXPOSURE]: The documentation uses placeholder strings such as sk-lf-prod-chatbot-... and https://langfuse.your-domain.com. While these triggered automated alerts, they are standard documentation placeholders and do not represent actual sensitive data exposure.
Recommendations
  • Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 9, 2026, 03:46 AM