linkfox-shopee-product-detail

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

• Third-party content exposure detected (medium risk: 0.30). Runtime path scripts/shopee_product_detail.py → call_api(params) sends user-supplied productUrl into POST /shopee/product/detail, and the agent ingests outsider-authored free text from the Shopee listing page response (e.g., name, brand, description, attributes, etc.).

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I inspected the skill files for literal high-entropy values that could be active credentials. I found a base64url-like string hardcoded as the default LINKFOX_LOGIN_FIXED_UID value in scripts/onboarding.py:
• It is a non-trivial, random-looking base64 string ("eyJhX2lkIjoiNmEyMmM4YjA1YmM5MTZhIiwiZF9pZCI6IiJ9"), not a placeholder like YOUR_API_KEY.
• The code comment and usage indicate it is a fixed UID header used in the login chain ("登录链路固定 uid header（实测可用）") and it is sent as the "uid" header when no user token is present. That suggests it is usable in requests (i.e., functionally meaningful), so it meets the definition of a secret/credential.
• I did not flag environment variable names, sample API_KEY placeholders, or simple example passwords (none present) because they are explicitly documentation placeholders or non-sensitive defaults.

Therefore I flagged the hardcoded fixed UID value as a potential active credential.