Productivity

AI agent skill implementing Google's TurboQuant compression algorithm (ICLR 2026) — 6x KV cache memory reduction, 8x speedup, zero accuracy loss. Compatible with Claude Code, Codex CLI, and all Agent Skills-compatible tools.

16 production-hardened patterns for OpenClaw agents. WAL, anti-hallucination, ambiguity gate, compaction survival, QA gates, and multi-agent handoffs.

FormCraft series — 对话驱动的办公格式化 Skill，把粗糙输入变成专业交付物

🛠️ CLI + MCP server for installing AI agent skills, workflows, and MCP configs. Your personal toolkit for AI-powered development.

Generate same-day Codex work reports grouped by project, combining sessions, commits, and unstaged changes.

Wrap up the current coding session into a concise summary of outcomes, verification, risks, and next actions.

Turn one day of git commits into grouped Chinese daily work summaries.

A collection of agent skills that give AI agents structured, reusable workflows for tackling complex problems.

A manual-first agent harness for leading Codex, Claude Code, and Antigravity with clearer requests, safer execution, visible verification, and operator-focused workflows.

Structured context management framework for LLM agents. Implements Git-like operations (COMMIT, BRANCH, MERGE) to manage long-horizon agent memory.

File-based lightweight work selection and tracking skill for agents

Codex skill for choosing the best local skill stack before starting a project.

Onboard and verify OpenClaw model providers from API keys, keep only working models, and sync chat-visible model allowlists safely.

Train a forgetful OpenClaw shrimp into a usable work partner

面向 OpenClaw 工作区的互联网研究、项目评估与学习蒸馏技能

Build Dify workflows from business requirements with design, YAML generation, review, refactor, validation, nanny-mode guidance, and URL-driven app recognition.

Human-in-the-loop skill for AI agents — create tasks for humans to review, approve, or complete. REST API + MCP.