lead-qualification

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md: Step 1.5/Step 3 ingests enriched lead profile text returned from Apify (outsider-authored web-scraped LinkedIn content) into the LLM context via the “self-contained context package” that includes the batch lead rows.

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

• Potentially malicious external URL detected (high risk: 0.90). The enrichment script calls the Gooseworks proxy at https://app.gooseworks.ai (BASE_URL) at runtime to start Apify actor runs (executes remote scraping code), so the skill will execute remote code via that external endpoint.