workos-authkit-base

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill explicitly directs the agent to "WebFetch" the SDK README from public sources (npmjs.com or GitHub), which are user-authored, untrusted third-party pages the agent will read and use as authoritative guidance, exposing it to indirect prompt injection.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly requires a runtime "WebFetch" of the SDK README from external hosts (e.g., https://www.npmjs.com/package/{sdk-package-name} or a GitHub README/raw URL like https://raw.githubusercontent.com/.../README.md), and that fetched README is used as the authoritative source of instructions for implementation, so external content would directly control the agent's prompts/behavior.