article-illustrator

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

• Third-party content exposure detected (medium risk: 0.65). Step 2 “Analyze content” and Step 5 “Generate images” require reading the article’s text (positions/labels/numbers/quotes) and then embedding that article-derived free text into prompt files/LLM context; if the article content is authored by someone other than the operating user, it is outsider-authored free text.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)