Pinned Loading

1. wiki-recall wiki-recall Public

   Compiled knowledge meets layered recall. Karpathy wiki + MemPalace memory layers. 5-layer architecture, ~550 tokens to wake up. 1,060 tests.

   TypeScript 2

2. a2a-crews a2a-crews Public

   Turn one command into a team of AI agents. Built on Google's A2A protocol.

   TypeScript 2

3. microsoft/conductor microsoft/conductor Public

   A CLI tool for defining and running multi-agent workflows with the GitHub Copilot SDK and Anthropic Agents SDK.

   Python 62 9

4. rag-a2a rag-a2a Public

   Production-grade RAG pipeline exposed as an A2A agent. Hybrid search, reranking, evaluation, streaming — the knowledge layer for multi-agent systems.

   TypeScript 1

5. agent-traps-lab agent-traps-lab Public

   Empirical testbed for DeepMind's AI Agent Traps paper — 22 adversarial scenarios, 4 models, 3,480 experiments

   TypeScript 1

6. ag-ui-crews ag-ui-crews Public

   Real-time dashboard for a2a-crews — built on the AG-UI protocol. Watch your AI agent crews plan, execute, and deliver in a live web UI.

   TypeScript 1