cortex-training-rl

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

• Third-party content exposure detected (medium risk: 0.30). The workflow monitors and consumes external task directories containing untrusted user-authored files (instruction.md and tests/test.sh) via Harbor.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill instructs cloning repositories from GitHub (Snowflake-AI-Research/Arctic-Platform and NovaSky-AI/SkyRL) during setup, constituting an operational dependency from an external source.