paper-writing-bench

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md workflow and references/bench-overview.md explicitly require ingesting paper PDFs or extracted markdown from public repositories (e.g., OpenReview and the CVF Open Access repository) and passing the [PAPER CONTENT] into the verbatim prompts (references/*.md), so arbitrary third‑party paper content is read and directly drives LLM outputs.