agent-package-manager

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

• Third-party content exposure detected (low risk: 0.10). The agent package manager skill installs remote packages and MCP servers from git repositories and registries, where an outsider could potentially author malicious package content or instruction files.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs downloading and executing remote scripts from aka.ms redirects during bootstrap/setup, which presents a weak/unknown provenance operational dependency.