Oracle ADK Expert
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFENO_CODEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists entirely of documentation, code snippets, and architectural guidelines. It does not include any executable scripts or binary files, which inherently minimizes the direct risk of malicious code execution.
- [DATA_EXFILTRATION]: The documentation includes code examples that reference the OCI configuration file at
~/.oci/config. While this is the standard authentication method for the OCI SDK, this file is a sensitive path containing credentials that developers must protect from unauthorized access. - [PROMPT_INJECTION]: The skill outlines patterns for agents to ingest untrusted data from user queries to drive actions like database interaction or tool execution, creating a surface for indirect prompt injection.
- Ingestion points: User input enters the agent context via
conversation.send()andorchestrator(user_query)as documented inSKILL.md. - Boundary markers: The provided code snippets do not demonstrate the use of delimiters or instructions to ignore instructions embedded within the user-provided data.
- Capability inventory: The documented patterns feature capabilities such as database access through
SQLTooland integration with OCI services like object storage and autonomous databases. - Sanitization: The examples do not include input validation, sanitization, or filtering of the external content before it is processed by the agent.
Audit Metadata