Pinned Loading

1. llm-benchmark llm-benchmark Public

   大模型横评：多任务提示词与各模型表现对比

   HTML 5

2. skills-papa skills-papa Public

   One inventory, every agent. Manage Agent Skills across Codex, Claude Code, OpenCode & universal roots: dry-run by default, atomic upgrades, rollback, centralized backups. 一份清单，管好所有 agent 的技能生命周期。

   Python 1

3. local-workbench local-workbench Public

   本地工作工具箱：跨 AI CLI 的历史回溯、本机运行台与交付核验台。Local-first workbench: cross-tool history search, machine monitor, delivery evidence. Pure Python stdlib, no cloud.

   Python