managing-oracle-oci-database

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses oci CLI commands for infrastructure management. All commands are limited to read-only operations such as list, get, and summarize-metrics-data. This is appropriate for its stated purpose of database analysis and management.
  • [COMMAND_EXECUTION]: Shell script snippets utilize the standard OCI_COMPARTMENT_ID environment variable for scoping, which is a standard practice for OCI automation.
  • [PROMPT_INJECTION]: The skill includes strong instructional guardrails under 'Safety Rules' and 'Anti-Hallucination Rules', specifically instructing the agent to never stop or terminate resources without explicit user confirmation and to never assume resource names.
  • [DATA_EXFILTRATION]: No network operations or external data transfer patterns were detected. Output is limited to the console using --output table and pipe to head for brevity.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 06:00 AM
Security Audit — agent-trust-hub — managing-oracle-oci-database