ai-for-science-boltz2

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs cloning and installing a public GitHub repository ("git clone https://github.com/jwohlwend/boltz.git" and "pip install -e .") and using weights from community sources, so the agent/workflow will fetch and execute untrusted, user-provided third‑party code that can alter runtime behavior.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs to run "git clone https://github.com/jwohlwend/boltz.git" and later "pip install -e .", which fetches remote code at runtime and installs/executes it from that repository (https://github.com/jwohlwend/boltz.git), so this external URL is a runtime dependency that can execute remote code.