performing-service-account-audit
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/agent.pyutilizes thesubprocess.runfunction to execute local command-line tools for discovery. Specifically, it invokespowershellto perform Active Directory user queries and theawsCLI to enumerate IAM users and access keys. - [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes metadata from external environments (AD and AWS), which represents a surface for indirect instructions potentially embedded in system data.
- Ingestion points: CLI tool output captured and parsed in
scripts/agent.py. - Boundary markers: No explicit delimiters are used to isolate untrusted metadata from the agent's execution logic.
- Capability inventory: The skill executes system shells and management utilities via
subprocess. - Sanitization: Ingested account properties (e.g., names, descriptions) are used in reports without explicit validation or escaping.
Audit Metadata