linkfox-temu-ads-us

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill includes examples and instructions that place API keys/access tokens directly into command-line arguments and JSON (e.g., export and inline "accessToken"), which requires the LLM to handle or emit secret values verbatim.

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

• Third-party content exposure detected (medium risk: 0.30). Skill “linkfox-temu-ads-us” at runtime reads and parses free-form JSON from the caller via CLI argument (scripts/us_ads_* → _us_ads_script.run_cli → _temu_common.load_json_arg → emit_result), and then forwards it through the Temu gateway (POST https://tool-gateway.linkfox.com/temu/proxy) as request fields.

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 tokens that could provide access. Most candidate strings are clearly placeholders (e.g. "", "TOKEN", "PASTE_TOKEN_HERE"), environment variable names (LINKFOXAGENT_API_KEY, TEMU_TOKEN_STORE_PATH, etc.), or opaque Partner/menu IDs (hex-like sub_menu_code values) which are not credentials and are ignored per the rules.

However, scripts/onboarding.py contains a hardcoded base64url-looking value assigned to _LOGIN_FIXED_UID: "eyJhX2lkIjoiNmEyMmM4YjA1YmM5MTZhIiwiZF9pZCI6IiJ9". The surrounding comment says this is a fixed uid header for the login flow and marks it "实测可用" (tested/works). That indicates it is a high-entropy, literal value used in an authentication header and therefore functions as a real credential-like secret rather than a documentation placeholder. Per the definition, this should be flagged.