install-hve-core-rpiv

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

• Third-party content exposure detected (high risk: 0.85). Runtime path: the skill uses gh api/git clone to fetch upstream HVE Core prompt/agent files from https://github.com/microsoft/hve-core (public web content), then reads that fetched free-text into the installer to render generated SKILL.md files—this is outsider-authored content.

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

• Potentially malicious external URL detected (high risk: 1.00). The installer fetches upstream prompt and agent files at runtime from the HVE Core repository (e.g., via gh api "repos/microsoft/hve-core/contents/...?ref=…" and/or git clone https://github.com/microsoft/hve-core), and those fetched files are directly used to generate the skill prompts, so the external repo URL is a required runtime dependency that controls agent prompts.