The Agent Skills Directory

[SAFE]: The skill provides legitimate scaffolding templates for the Open Reward Standard framework without any detected malicious patterns or obfuscation.
[EXTERNAL_DOWNLOADS]: The skill instructs the agent to include the openreward package from PyPI and references its official repository at github.com/Talc-AI/OpenReward. It also correctly warns against using the incorrect ors-sdk package name, which helps prevent dependency confusion attacks.
[COMMAND_EXECUTION]: The provided templates include standard development commands for running a Python server (python server.py) and validating endpoints using curl. These are appropriate for the skill's stated purpose of environment development.
[DATA_EXFILTRATION]: The instructions explicitly advise users to handle sensitive credentials like E2B_API_KEY using platform-specific secret management (Hugging Face Space secrets) instead of hardcoding them, aligning with security best practices.

generate-ors-env