octen-design

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

• Third-party content exposure detected (medium risk: 0.30). The octen-design skill reads design reference text at runtime from the octen.ai image-search API response, specifically html_snippet/summary fields saved to *.html and later consumed as design material (and written into results.json as octen_refs).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime Python script posts to the image-search endpoint https://api.octen.ai/image-search (scripts/search.py) and consumes returned "summary" and "html_snippet" data that are written to results.json and presented to the model to drive generated design content, so remote content can directly control prompts.