The Agent Skills Directory

[COMMAND_EXECUTION]: The skill extensively uses the Bash tool to execute aws CLI commands defined in various reference files (e.g., references/programmatic-checks/fcd-04-auditing.md). These commands are used to discover and analyze the configuration of the AWS environment being assessed.
[DYNAMIC_EXECUTION]: The skill performs dynamic execution by piping AWS CLI output into embedded Python scripts (using python3 -c or heredocs) to parse complex JSON structures, as seen in references/programmatic-checks/fcd-03-disclosure-controls.md and fcd-06-authentication.md. It also executes a local report generation script scripts/generate-html-report.py.
[INDIRECT_PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting untrusted configuration data from the AWS environment. This risk is mitigated by the skill's mandatory read-only credential gate defined in references/credential-boundary.md and the use of structured parsing for all external data ingestion.

ferpa-reviewer