linkfox-shopee-store-public

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

• Third-party content exposure detected (medium risk: 0.30). 该 skill 运行时通过 scripts/public_api.py/_public_api_runner.run_public_api() 将用户提供的 JSON（含 api、各字段）发起 POST /shopee/developerProxy，并在 merge_shopee_body() 中把网关返回的 body 解析为 JSON 后写入文件与 stdout摘要/全量，因此会消耗“外部用户文本输入”的运行时内容（间接提示注入面：用户可控字段会进入请求体/查询字符串并影响后续解析展示）。

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

• Secret detected (high risk: 1.00). I scanned the repo for literal high-entropy values that could be usable credentials.

Flagged:

• scripts/onboarding.py contains a hardcoded base64-like UID string assigned to _LOGIN_FIXED_UID (the code comment even says the fixed uid header is "实测可用" / practically usable). This is a literal JWT-like/base64 value embedded as a default rather than only a placeholder, so it may be a real, usable header/token and meets the "literal value that provides access" risk.

Ignored (not flagged) examples and false positives:

• Placeholders such as "", "xxx", or environment variable names like LINKFOX_AGENT_API_KEY / LINKFOXAGENT_API_KEY — these are documentation placeholders or env var names and intentionally ignored per rules.
• Sample/simple strings in examples (e.g., "67890", "xxx") and other numeric constants (device id, ports) — low entropy or example values not considered secrets.