setup-openflow-privatelink

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were identified. The skill provides legitimate technical documentation and workflow instructions for infrastructure configuration.
  • [COMMAND_EXECUTION]: The skill requires the use of high-privilege accounts (Snowflake ACCOUNTADMIN and AWS administrative permissions) to execute system functions and modify network permissions. This level of access is appropriate and documented for the specific task of provisioning PrivateLink endpoints.
  • [PROMPT_INJECTION]: The skill demonstrates an indirect prompt injection surface because it interpolates user-supplied parameters (such as <endpoint_service_name> and <nlb_dns_name>) into SQL and Bash commands. While this is standard for setup templates, it represents a surface where untrusted input could influence command execution.
  • Ingestion points: User-provided variable placeholders in SKILL.md and TROUBLESHOOTING.md.
  • Boundary markers: None identified.
  • Capability inventory: Uses snowflake_sql_execute and Bash to perform network configuration and resource provisioning.
  • Sanitization: None specified; relies on the agent or user to provide valid identifiers.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 03:06 AM
Security Audit — agent-trust-hub — setup-openflow-privatelink