openclaw-rl-training

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

• Third-party content exposure detected (medium risk: 0.30). The workflow documents training an agent on captured live user conversations, which contain untrusted outsider-authored free text ingested into the reinforcement learning pipeline.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill references an external GitHub repository (https://github.com/Gen-Verse/OpenClaw-RL) in its installation instructions, which is cloned and executed at runtime, establishing a common/reputable unofficial operational dependency.