oci-services-expert
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches pricing data from official Oracle domains (oracle.com) and searches for repository information via the GitHub API targeting verified Oracle organizations (oracle-devrel, oracle-quickstart, oracle-samples). These are well-known technology services and trusted sources used for factual guidance.
- [COMMAND_EXECUTION]: Utilizes the GitHub CLI (gh api) to programmatically search for official reference architectures. The commands are specifically restricted to discovering Oracle-maintained content.
- [PROMPT_INJECTION]: The skill processes data from external web sources, creating a potential surface for indirect prompt injection. However, the risk is negligible as the sources are authoritative vendor domains and the tool capabilities are limited to information retrieval.
- Ingestion points: External data enters through WebFetch (oracle.com) and gh api (github.com).
- Boundary markers: None explicitly defined for the fetched content.
- Capability inventory: Informational tool use (GitHub search, web retrieval); no file-system writing or arbitrary command execution detected.
- Sanitization: Standard tool-based retrieval; no specific sanitization logic is described in the skill markdown.
Audit Metadata