The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Heroku CLI to perform read-only discovery of account resources.
Evidence: discover-live-capture.md contains a whitelist of heroku commands such as apps, ps, addons, and spaces.
Context: The skill includes a dedicated security contract that restricts execution to LIST/INFO commands and prohibits the capture of secrets like API tokens or configuration values.
[EXTERNAL_DOWNLOADS]: The skill fetches infrastructure resources and survey forms from official Amazon and AWS domains.
Evidence: https://truststore.pki.rds.amazonaws.com/global/global-bundle.pem (RDS CA bundle) and https://aws.github.io/eks-charts (Helm charts).
[INDIRECT_PROMPT_INJECTION]: The skill processes various untrusted files from the local workspace which serves as an indirect injection surface.
Ingestion points: discover-terraform.md (Terraform files), discover-billing.md (Billing exports), and discover-live.md (Procfiles).
Boundary markers: Absent. The agent relies on internal logic to parse these formats.
Capability inventory: Includes file reading, writing, and shell command execution.
Sanitization: The skill performs manual redaction of sensitive keys and values during the discovery process.
[DYNAMIC_EXECUTION]: The skill runs a local script to validate the integrity of generated migration reports.
Evidence: generate-assemble.md calls python3 "$PLUGIN_ROOT/scripts/validate-heroku-migration-report.py".

heroku-to-aws