The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on executing the AWS CLI (aws) and jq to audit cloud resources. This is the primary intended purpose of the skill.
Evidence: Found across all scripts in the scripts/ directory, used for read-only inventory (describe-regions, describe-instances) and gated mutations (delete-volume, modify-volume).
[EXTERNAL_DOWNLOADS]: The HTML dashboard template references a JavaScript library from an external CDN.
Evidence: assets/dashboard-template.html includes <script src="https://cdn.jsdelivr.net/npm/chart.js@4.4.1/dist/chart.umd.min.js"></script>. This targets a well-known service (jsDelivr) for a standard visualization library.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the AWS environment, such as resource tags and names, which are then interpolated into reports and the HTML dashboard.
Ingestion points: AWS CLI command outputs (e.g., aws ec2 describe-volumes --query 'Volumes[0].Tags') processed in SKILL.md and scripts.
Boundary markers: The skill uses defined Artifact templates and a JSON schema (findings.json) to structure output. references/output-and-reporting.md defines a strict output contract.
Capability inventory: Shell command execution via aws CLI, file writes for logs, artifacts, and dashboard generation.
Sanitization: The assets/dashboard-template.html includes an esc() function to escape data before rendering it in the browser.

aws-cost-audit