linkfox-jiimore-get-niche-info-by-asin

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

• Third-party content exposure detected (low risk: 0.10). Skill “Jiimore Niche Info by ASIN” only ingests outsider-authored text if the user supplies it as the POST JSON parameters (notably asin), and then the script merely forwards that to the first-party gateway and prints/summarizes the gateway response, without monitoring any outsider message stream or requiring the agent to fetch arbitrary outsider content.

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

• Secret detected (high risk: 1.00). I flagged one literal that appears to be a high-entropy token-like value embedded in code. Specifically, scripts/onboarding.py contains a base64url string assigned to _LOGIN_FIXED_UID which is used as a uid header in the login flow ("实测可用" comment) — this is a static, token-like value that could be used by the code to interact with authentication endpoints, so it meets the definition of a potential secret.

I ignored other values:

• Environment variable names like LINKFOX_AGENT_API_KEY / LINKFOXAGENT_API_KEY are just names/placeholders (no secret literals present).
• Example ASINs and other sample data in SKILL.md and references are non-secrets (not high-entropy credentials).