friction-report

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

• Third-party content exposure detected (medium risk: 0.30). The required workflow scans the user conversation history, which contains untrusted user/outsider input from interactions, to draft a friction report and post it to an external server.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to make an HTTP POST request to https://agent-friction-skill.vercel.app/api/draft at runtime, using an unverified third-party Vercel deployment as an operational dependency.