visual-prior-art-survey

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow deep-fetches and reads angle sources from the tool’s own source registry (e.g., b5 uses GitHub-docs https://github.com and b4/a1 use other index-style sources), so outsider-authored free text from those sites can be ingested at runtime when the agent selects and fetches candidates.