skills/mukul975/anthropic-cybersecurity-skills/implementing-privileged-access-management-with-cyberark/Gen Agent Trust Hub
implementing-privileged-access-management-with-cyberark
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides legitimate tools for CyberArk PAM administration and auditing. The code adheres to expected security patterns for interacting with the CyberArk REST API.
- [INDIRECT_PROMPT_INJECTION]: The audit scripts ingest data from external CyberArk API endpoints (Safes, Accounts, Platforms). This creates a vulnerability surface for indirect prompt injection if malicious data is present in the vault.
- Ingestion points:
scripts/agent.py(line 67, 88, 125) - Boundary markers: Absent from audit output
- Capability inventory: Performs authenticated network operations (GET/POST) via the requests library
- Sanitization: API data is printed directly without sanitization
Audit Metadata