oraclecloud-migration-deep-dive

Pass

Audited by Gen Agent Trust Hub on Sep 9, 2026

Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [DATA_EXFILTRATION]: The skill accesses the local OCI configuration file to authenticate migration commands and SDK calls. \n
  • Evidence: The skill uses oci.config.from_file("~/.oci/config") to initialize the OCI Python SDK clients in SKILL.md.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes configuration data exported from AWS/Azure CLI, which could potentially contain malicious content if the source environment is compromised.\n
  • Ingestion points: aws-vpc.json, aws-subnets.json, and aws-routes.json files in SKILL.md.\n
  • Boundary markers: Absent. No specific delimiters are used to separate untrusted data from instructions.\n
  • Capability inventory: File system writes, VCN creation, IAM policy management, and Object Storage uploads via OCI CLI and SDK.\n
  • Sanitization: Absent. The scripts directly process the exported JSON without validation steps.\n- [EXTERNAL_DOWNLOADS]: The skill directs the user to install the official OCI Python SDK from the standard package registry.\n
  • Evidence: pip install oci instruction in the Prerequisites section of SKILL.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 9, 2026, 03:49 AM
Security Audit — agent-trust-hub — oraclecloud-migration-deep-dive