baoyu-xhs-images

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow ingests outsider-authored free text when a user supplies direct content (either via a pasted article in the “Direct content input” mode or by providing a source file path), and then reads that saved “source.md/source-{slug}.{ext}” content for Step 1 analysis.