Adaptive PPO Exploration via Reward History
100 / 100 · Low risk
Fine to try.
Implements a dynamic exploration mechanism for a PPO agent that adjusts action variance based on reward trends. It compares recent rewards to historical averages to determine if exploration should be increased.
What this skill can do
We found nothing in the files we read that matches our rules. That does not prove the skill is safe: see the limits of static scanning.
Other copies in this repository
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'adaptive-ppo-exploration-via-reward-history' ] || [ -L ~/.agents/skills/'adaptive-ppo-exploration-via-reward-history' ]; then echo 'A skill folder named adaptive-ppo-exploration-via-reward-history already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/ECNU-ICALK/AutoSkill' "$d" && git -C "$d" checkout --quiet '94c47ca488d4ba4117d20272e66d49b9877e68cf' -- 'SkillBank/ConvSkill/english_gpt4_8/adaptive-ppo-exploration-via-reward-history' && mkdir -p ~/.agents/skills && cp -R "$d"/'SkillBank/ConvSkill/english_gpt4_8/adaptive-ppo-exploration-via-reward-history' ~/.agents/skills/'adaptive-ppo-exploration-via-reward-history' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'adaptive-ppo-exploration-via-reward-history' ~/.claude/skills/'adaptive-ppo-exploration-via-reward-history' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
[](https://zalavijaysinh.github.io/skilltrust/repos/ECNU-ICALK/AutoSkill/)
Score history
| Date | Version | Score |
|---|---|---|
| 2026-11-08 | 94c47ca | 100 |
| 2026-11-07 | 94c47ca | 100 |