dingtalk-adoc-reader

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill requires using full browser session cookies and XSRF tokens (copied or returned by QR login) and embedding them verbatim in HTTP headers/requests, which forces the LLM to handle secrets directly and risks exfiltration.

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

• Third-party content exposure detected (high risk: 0.90). This skill programmatically fetches and parses user/knowledge-base documents from the third-party web portal https://alidocs.dingtalk.com (see SKILL.md and client.py: get_dentry_children, get_dentry_info, get_document_data, export_to_pdf) and then runs extract_text_from_alidocs and export/download flows that interpret that untrusted, user-generated content and drive further actions, enabling indirect prompt-injection risk.