windmill

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly shows fetching public third‑party content (curl https://raw.githubusercontent.com/windmill-labs/windmill/main/docker-compose.yml and example scripts using requests.get("https://api.example.com/data") and a "Fetch data from API" workflow step), which indicates the platform ingests external/untrusted API/web data that can influence branching and follow-up actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's self-host instructions fetch and execute remote code at runtime — e.g., pulling/running the container image ghcr.io/windmill-labs/windmill:main and curling https://raw.githubusercontent.com/windmill-labs/windmill/main/docker-compose.yml (then running docker compose) — so external content is fetched and executed and is a required dependency.