setting-up-cloudtrail-multi-region
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- Indirect Prompt Injection Surface: The skill uses user-provided parameters like names and region identifiers directly within command strings. This pattern creates a surface where specifically crafted inputs could potentially influence the execution of commands or the structure of generated policies. * Ingestion points: Parameters such as trail_name and s3_bucket_name defined in references/cloudtrail-multi-region-setup.md. * Boundary markers: The skill does not provide specific instructions or delimiters to isolate user-supplied parameters from the command logic. * Capability inventory: The skill uses call_aws to perform s3api create-bucket, logs create-log-group, iam create-role, and cloudtrail create-trail across the steps in references/cloudtrail-multi-region-setup.md. * Sanitization: No input validation, escaping, or filtering of parameters is specified in the instruction set.
- Command Execution: The skill is designed to perform sequence-based resource configuration through command-line tools. This is a functional requirement for its stated purpose.
Audit Metadata