linkfox-kalodata-tiktok-livestream

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

• Third-party content exposure detected (low risk: 0.10). The Kalodata skill’s runtime path POST /kalodata/livestream/rank and then POST /kalodata/livestream/detail ingests outsider-authored free text only indirectly via TikTok livestream titles/handles returned by the ranking/detail endpoints, not from any user-provided text that the workflow reads verbatim into the LLM context.

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

• Secret detected (high risk: 1.00). I scanned the files for literal credentials. Most potential sensitive items are environment variable names or placeholders (e.g., LINKFOX_AGENT_API_KEY, LINKFOXAGENT_API_KEY, ${LINKFOX_TOOL_GATEWAY}) or example IDs (e.g., 7661409374878878494) which are documentation/example values and not secrets per your rules.

However, scripts/onboarding.py contains a hardcoded base64url-like string assigned to _LOGIN_FIXED_UID:

• It is a non-trivial, encoded string (not a simple placeholder) and is explicitly used as a "登录链路固定 uid header（实测可用）" which implies it functions as an authentication header value. This fits the definition of a literal value that could provide access/identity and is therefore flagged as a secret.

No other high-entropy API keys, private keys, or similar literals were found. Standard examples, numeric IDs, and environment-variable placeholders were ignored per the given rules.