alphaxiv

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow (Steps 2–4 in SKILL.md) explicitly fetches and parses untrusted public content from alphaxiv.org (e.g., https://alphaxiv.org/overview/{PAPER_ID}.md and https://alphaxiv.org/abs/{PAPER_ID}.md) and arXiv source archives (https://arxiv.org/src/{PAPER_ID}), and the agent is expected to read and act on that content to answer questions and decide follow-up actions, which could allow indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly fetches runtime content from https://alphaxiv.org/overview/{PAPER_ID}.md (and falls back to https://alphaxiv.org/abs/{PAPER_ID}.md), which are described as "structured, LLM-optimized" reports that are used directly as the agent's input/instructions, so external content controls prompts and is a required dependency.