oraclecloud-incident-runbook

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFECREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [CREDENTIALS_UNSAFE]: The skill accesses the OCI configuration file at ~/.oci/config. This file contains sensitive credentials, including API keys, fingerprints, and tenancy OCIDs, which are required for the OCI SDK and CLI to function. This access is consistent with the skill's stated purpose of managing cloud infrastructure.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from OCI API endpoints to determine service health and instance lifecycle states.
  • Ingestion points: API responses from identity.list_regions, compute.list_instances, network.list_vcns, and oci compute instance get commands (found in SKILL.md).
  • Boundary markers: None identified; the skill compares API output strings directly (e.g., checking if status equals "RUNNING").
  • Capability inventory: Includes high-impact actions such as oci compute instance action --action RESET and compute.launch_instance for failover.
  • Sanitization: The skill relies on standard OCI CLI query filtering (--query) but does not perform additional validation on the returned strings before using them in conditional logic.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:46 AM
Security Audit — agent-trust-hub — oraclecloud-incident-runbook