adk-verify-snippets

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

• Third-party content exposure detected (medium risk: 0.30). In the required runtime workflow, verify_md.py reads an outsider-authored Markdown file from <path_to_markdown_file.md>, extracts ```python blocks, writes them to temp .py files, and then run.py executes/loads any discovered module-level Agent/Workflow/App—so attacker-authored text can be ingested indirectly via the executed snippet code.