linkedin-influencer-discovery

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The script reads an API key from the environment and deliberately sends it (as "token") to a non-Apify endpoint (GOOSEWORKS_API_BASE defaulting to app.gooseworks.ai) and then fetches datasets including contact info — behavior consistent with credential exfiltration / unauthorized data forwarding.

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

• Third-party content exposure detected (high risk: 0.75). The required workflow calls an Apify actor at runtime and ingests the returned dataset items (including outsider-authored influencer profile text like biography) into the script’s output, which can then be fed into the agent/LLM context as free text.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The script calls the runtime API at https://app.gooseworks.ai to start and poll an Apify actor run (remote code execution) and the skill requires the GOOSEWORKS_API_KEY/Apify token to operate, so this external endpoint controls execution at runtime.