upwork-apply

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

• Third-party content exposure detected (high risk: 0.90). The skill scrapes public Upwork job listings via Apify (scripts/upwork_apify_scraper.py and SKILL.md) and feeds untrusted, user-generated job titles/descriptions directly into the proposal-generation prompts in scripts/upwork_proposal_generator.py (passed to Claude), so third-party content can influence LLM outputs and subsequent actions (Google Doc creation, sheet updates).

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

• Potentially malicious external URL detected (high risk: 0.90). The skill makes runtime calls to Apify (e.g. https://api.apify.com/v2/acts/upwork-vibe~upwork-job-scraper/runs?token=... and the corresponding https://api.apify.com/v2/datasets/{dataset_id}/items?token=...) to fetch Upwork job data which is then injected directly into Anthropic/Claude prompts to generate proposals, so external content fetched at runtime directly controls model prompts and is a required dependency.