The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches documentation, examples, and supplementary skills from vendor-owned domains and repositories including superwall.com and github.com/superwall. These are legitimate resources provided by the author to support the framework's functionality.
[REMOTE_CODE_EXECUTION]: Instructions include the installation of additional specialized skills (e.g., mobile-design, onboarding-questions) using npx skills add from the vendor's official GitHub repository. This is an intended mechanism for extending the framework's capabilities.
[COMMAND_EXECUTION]: The skill documents the use of the superwall CLI for standard development lifecycle tasks such as create, dev, push, promote, and pull. These commands are essential to the primary purpose of the framework.
[DATA_EXFILTRATION]: A feedback mechanism (superwall feedback) is documented which allows users to send bug reports and logs to the vendor. The instructions explicitly advise users to strip secrets and customer data before submission.
[INDIRECT_PROMPT_INJECTION]: The skill features migration playbooks that ingest and process external data, specifically native application source code (Swift, Kotlin, etc.) and dashboard JSON. While this creates an attack surface for indirect injection, the instructions emphasize schema validation, typed parsers, and human review gates to mitigate risks.

superwall-framework