Agent Skills Hub

local-skills-mcp-guide

Repository implementation guide for local-skills-mcp. Provides technical documentation on MCP tool handlers, skill loading, aggregation logic, and project structure for developers.

hs

Pre-execution security guardrails for AI agents. Validates shell commands and file reads against 400+ security patterns to block destructive operations, credential theft, and unauthorized system access.

pytorch-lightning

PyTorch Lightning skill for scalable deep learning: automates model training, multi-GPU orchestration, data pipelines, and distributed training strategies like DDP, FSDP, and DeepSpeed.

explain-concepts

Explains complex concepts using master teaching frameworks like Feynman, Socratic, and Cognitive Load theory to ensure deep, clear understanding.

upstash-workflow-js

Build durable, reliable serverless workflows using the Upstash Workflow SDK. Define endpoints, manage complex execution steps, and integrate with QStash for automatic retries and state management.

freeride

Manages free AI models from OpenRouter for OpenClaw. Ranks models by quality, configures fallbacks for rate-limit handling, and updates openclaw.json automatically.

familiar-add-tool

Scaffold and register new sensor, actuator, or service tools for familiar-ai, automating file creation and boilerplate integration in agent.py and config.py.

htmx-rust

Build interactive, hypermedia-driven web applications using Rust, Axum, and HTMX for dynamic, real-time UI updates without complex JavaScript frameworks.

dispatching-parallel-agents

Parallelize independent debugging or development tasks by delegating to specialized subagents with isolated context.

vm-dev

Virtual machine development expert focusing on bytecode design, stack-based/register-based VM implementation, memory management, and garbage collection.

mcp-research

Retrieve current, source-backed technical information using MCP tools to resolve queries about libraries, APIs, SDKs, and evolving tech ecosystems.

massive-context-mcp

Process massive files and large codebases (10M+ tokens) by recursively chunking, sub-querying, and aggregating results to overcome LLM context limits.