mcp-client
mcp-client is a universal client for connecting to MCP servers, allowing for efficient browser session management and multi-step operations.
Browse and install thousands of AI Agent skills in the Killer-Skills directory. Supports Claude Code, Windsurf, Cursor, and more.
mcp-client is a universal client for connecting to MCP servers, allowing for efficient browser session management and multi-step operations.
agentuity-cli-dev is a command-line interface tool for building and running development servers, utilizing the bunx command.
openspec-continue-change is a minimalistic AI API proxy that continues working on a change by creating the next artifact, using commands like openspec list --json.
Common is a library of utilities for AI agent skills using the Gemini API, offering features like API key management and Vertex AI support.
knowns.task.brainstorm is a task design skill that uses structured questioning to understand problems before proposing solutions, utilizing a phased approach for efficient design.
tfm-add-card is a skill that allows developers to add new cards to TFM by registering the card name, implementing the card, registering in manifest, and writing unit tests.
karpathy is a set of coding principles for writing clean, minimal, and purposeful code, focusing on clarity and simplicity in AI agent development.
workflow-to-skill is a process extraction tool that converts unstructured user operations into structured workflow maps, decision trees, and quality checkpoints.
spp-stats is a technical skill that assesses weekly gate metrics, including human and AI contributions, to provide actionable insights for AI agent optimization.
new-item is a Minecraft Bedrock add-on creation skill that automates the generation of item scaffolds for Mega Knights items.
claude-code is a specialized AI agent skill for autonomous coding, integrating subagents, agent skills, and MCP servers for efficient development workflows.
A Multi-Agent System (MAS) evaluation framework using PydanticAI that generates and evaluates scientific paper reviews through a three-tiered assessment approach: traditional metrics, LLM-as-a-Judge, and graph-based complexity analysis.