tooluniverse-custom-tool

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 tool implementations that perform HTTP fetches (e.g., the Python run() example using requests.get("https://my-api.example.com/search") and JSON config "endpoint" fields) and even references public URLs (e.g., certification.oshwa.org), meaning the skill instructs the agent to ingest arbitrary third‑party web API/content as part of its workflow which could materially influence subsequent behavior.