auditing-access-control
Pass
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses scoped Bash utilities (e.g., security:, scan:, audit:*) and standard commands like grep to identify access control definitions and permission grants in the target codebase.
- [DATA_EXFILTRATION]: Analysis is performed exclusively on local codebase and configuration files; no network tools, external URLs, or data exfiltration patterns were identified in the scripts or instructions.
- [PROMPT_INJECTION]: The skill processes untrusted external data such as IAM policies and source code as part of its auditing function. While this creates a surface for indirect prompt injection, the skill's logic is focused on legitimate security assessment and does not contain instructions to bypass agent safety filters or override system constraints.
Audit Metadata