pmem-style-extraction

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

• Third-party content exposure detected (high risk: 0.90). The skill's SKILL.md shows it will fetch arbitrary public webpages using the baoyu-url-to-markdown call (npx -y bun ${SKILL_URL_DIR}/scripts/main.ts -o /tmp/extracted_content.md) and then reads that untrusted page content into an LLM-driven feature-extraction workflow that directly influences conflict detection, profile creation, and storage decisions, creating a clear path for indirect prompt injection.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches arbitrary webpages at runtime using the baoyu-url-to-markdown flow (e.g. https://example.com/article) and injects the retrieved content into the LLM prompt ({CONTENT}), meaning an external URL can directly control the model input.