Agent Skills · AI Agent
Curated skills for Claude, Cursor and other AI agents. Copy one prompt to install.
Converts a research need into an executable Deep Research prompt using the 6D workflow: define, decompose, design, discover, digest, and deliver.
Design and optimize LLM prompts with system prompts, chain-of-thought, few-shot strategies, and evaluation loops to reduce instability and hallucination.
A Godot 4.3+ animation skill for choosing and driving AnimationPlayer, AnimationTree, and Tween to run keyframes, state blends, and code tweens.
A passive agent skill for opencode that isolates dev memory after confirmation, records project state, tasks, and architecture decisions, and restores context.
Explores, verifies, and packages website data extraction or operation needs into deployable Agent browser capabilities.
Logs agent errors, corrections, and feature requests, then promotes durable learnings into project memory.
Create, optimize, and distill high-quality prompts through structured questioning and reusable pattern extraction.
A local transparent proxy that intercepts OpenAI, Anthropic, and Ollama requests to reduce token usage via deduplication, filtering, and compression, with a savings dashboard.
A coach-style agent workflow for complex tasks: clarifies deliverables, runs a minimal loop, and uses controlled Hermes/ACP execution with Feishu notifications to ship usable results.
A Python wrapper for the Langflow AI workflow platform, covering visual builders, RAG pipelines, agents, and MCP server deployment.
Wraps LlamaIndex RAG to help add retrieval-augmented generation to automated workflows.
A TypeScript Cal.com wrapper for GitHub and automation-related integrations.