linkfox-tiktok-shop-product-detail

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

• Third-party content exposure detected (medium risk: 0.30). 技能在运行时读取用户提供的 productInput（TikTok 商品 URL 或 19 位商品 ID），并调用 POST /tiktok/shop/product/detail 获取上游返回的商品详情 JSON，该 JSON（含标题、卖家信息、媒体字段等）会被脚本保存到本地并在需要时打印到上下文/输出。

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

• Secret detected (high risk: 1.00). I scanned the skill files for literal high-entropy values that could act as credentials. Most strings are placeholders, environment-variable names, example product IDs, URLs, or defaults (these I ignored per the rules). I found one literal base64-like token assigned as a fixed UID header in the onboarding CLI:
• scripts/onboarding.py defines _LOGIN_FIXED_UID with a base64url string (not a placeholder). The code comment explicitly calls it a "fixed uid header (实测可用)" and the string is used as the fallback uid header for authenticated requests when no user_id is available. This is a literal token that appears reusable in network requests (i.e., not a simple example password or a harmless placeholder), so it meets the definition of a secret to flag.

No other high-entropy literal API keys, private keys, or similar credentials were present. Example keys and instructions that reference environment variables (e.g., LINKFOX_AGENT_API_KEY) are placeholders and were ignored.