linkfox-temu-price-global

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill's examples and run commands explicitly show embedding API tokens/accessTokens in command-line JSON (and an export placeholder), which would require an LLM to include secret values verbatim in generated commands or snippets — an exfiltration risk.

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

• Third-party content exposure detected (medium risk: 0.30). 该 skill 通过 CLI 入参 JSON 直接将用户/外部调用者提供的 params（“业务载荷”）拼入网关请求并发送到 scripts/temu_global_proxy.py 的 POST /temu/proxy（最终由上游 Temu 返回的字符串/字段也会被解析与输出），因此若有人能在运行时把“外人文本”放进该 JSON 中且传入会被当作自由文本字段的一部分，LLM 运行时会间接接触到未受约束的外部文本。

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

• Secret detected (high risk: 1.00). I scanned the repository for literal high-entropy values that could be used as credentials. The only literal that looks like a non-placeholder, high-entropy token is the base64url string hardcoded as the default UID header in scripts/onboarding.py:
• It is a compact base64url JWT-like segment (not a simple word or placeholder) and is used as a fixed uid header fallback (_LOGIN_FIXED_UID) for login requests. Although there is an environment variable override, the literal default is present in code and is used when no env var is set. A literal like this can act as an access header and therefore meets the definition of a secret to flag.

All other examples in the docs (e.g., "", "TOKEN", "PASTE_TOKEN", menu_code/sub_menu_code hex identifiers, simple passwords) are placeholders, doc identifiers, or low-sensitivity values and were ignored per the rules.