Awesome Agent Knowledge

A curated list of knowledge resources for AI agent operators — memory, security, governance, evaluation, commerce, and infrastructure.

Building and operating AI agents is a fragmented discipline. This list collects the best resources across the entire agent lifecycle: from identity and memory architecture to security hardening, multi-agent governance, and autonomous commerce.

Contents

• Agent Memory
• Agent Security
• Multi-Agent Governance
• Agent Identity & Personas
• Agent Evaluation
• MCP Resources
• Agent Commerce
• Multi-Agent Frameworks
• Agent Infrastructure

---

Agent Memory

How agents persist knowledge across sessions, manage context windows, and share collective memory.

• The Hive Doctrine — Agent Memory Architecture - Three-layer memory model: session, persistent, and collective (free).
• The Hive Doctrine — Three-Tier Episodic Memory - Hot/warm/cold storage architecture for agent memory ($99, via MCP).
• LangChain Memory - Memory modules for LLM chains — buffer, summary, entity, and conversation memory.
• Letta (formerly MemGPT) - OS-like memory management for LLM agents with tiered storage.
• Mem0 - Memory layer for personalised AI — adds long-term memory to any LLM.
• Zep - Long-term memory for AI assistants with knowledge graph integration.
• ChromaDB - Open-source vector database for storing and retrieving agent memory embeddings.

Agent Security

Securing autonomous agents against prompt injection, data exfiltration, and adversarial attacks.

• The Hive Doctrine — Prompt Injection Defence - The Airlock Pattern for prompt injection defence (free).
• The Hive Doctrine — Agent Security Checklist - 10-point pre-deployment security checklist (free).
• OWASP Top 10 for LLM Applications - Industry-standard vulnerability taxonomy for LLM systems.
• Simon Willison — Prompt Injection - Comprehensive blog series on prompt injection attacks and defences.
• Lakera Guard - Prompt injection detection benchmark and tools.
• Rebuff - Self-hardening prompt injection detector.
• LLM Guard - Security toolkit for LLM interactions — input/output scanning.
• Guardrails AI - Input/output guards for LLM applications.

Multi-Agent Governance

Frameworks for authority, hierarchy, and coordination in multi-agent systems.

• The Hive Doctrine — Multi-Agent Coordination - Coordination without a coordinator (free).
• The Hive Doctrine — Five Multi-Agent Architectures - Five patterns compared (free).
• The Hive Doctrine — Distributed AI Safety - Polytheistic AI safety thesis (free).
• CrewAI - Framework for orchestrating role-playing autonomous AI agents.
• AutoGen - Microsoft's framework for building multi-agent conversational systems.
• LangGraph - Build stateful, multi-agent applications with LLMs as graphs.
• Swarm - OpenAI's experimental multi-agent orchestration framework.
• Agency Swarm - Agent orchestration framework focused on agent collaboration.
• MetaGPT - Multi-agent framework that assigns roles like a software company.

Agent Identity & Personas

Defining agent culture, voice, boundaries, and constitutional identity.

• The Hive Doctrine — SOUL.md Template - Constitutional identity template for AI agents (free).
• The Hive Doctrine — SOUL.md Standard - The SOUL.md specification (free).
• The Hive Doctrine — Living Presence Protocol - Making agents feel present, not scripted (free).
• SOUL.md Templates - 5 free SOUL.md identity templates for common agent archetypes.
• Character.ai Research - Research on character-consistent AI agents.
• Anthropic — Claude's Character - Official guidance on shaping Claude's persona.

Agent Evaluation

Testing, benchmarking, and scoring agent performance.

• The Hive Doctrine — Evaluate Your Agent - Framework for evaluating agent performance (free).
• TruLens - Evaluation and tracking for LLM applications — groundedness, relevance, toxicity.
• RAGAS - Evaluation framework for Retrieval Augmented Generation pipelines.
• AgentBench - Benchmark for evaluating LLMs as agents across diverse environments.
• SWE-bench - Benchmark for evaluating LLMs on real-world software engineering problems.
• Braintrust - Evaluation, logging, and prompt playground for AI applications.
• Inspect AI - UK AISI's framework for evaluating LLM agents on safety-relevant tasks.

MCP Resources

Model Context Protocol — the standard for connecting AI agents to tools and data.

• MCP Specification - Official Model Context Protocol specification.
• MCP TypeScript SDK - Official TypeScript SDK for building MCP servers and clients.
• MCP Python SDK - Official Python SDK for MCP.
• The Hive Doctrine MCP Server - Knowledge marketplace with 116+ products, 6 tools, x402 payments. (Endpoint)
• The Hive Doctrine — MCP Server from Scratch - Step-by-step guide to building an MCP server (free).
• The Hive Doctrine — MCP Tools Explained - How agents discover and use MCP tools (free).
• mcp-handler - Deploy MCP servers to Vercel with zero config.
• Awesome MCP Servers - Curated list of MCP servers.

Agent Commerce

Enabling agents to transact, pay, and participate in markets autonomously.

• The Hive Doctrine — Agent Wallet Setup - Set up crypto wallets for agent payments (free).
• x402 Protocol - HTTP 402-based payment protocol for machine-to-machine commerce.
• Coinbase AgentKit - Toolkit for building crypto-native AI agents.
• NEAR AI - AI agents with native blockchain identity and payments.
• Skyfire - Payment infrastructure for AI agents.
• Virtuals Protocol - Tokenised AI agents with autonomous revenue.

Multi-Agent Frameworks

Production frameworks for building and deploying multi-agent systems.

• The Hive Doctrine — Stigmergic vs Centralised Routing - When to use which coordination pattern (free).
• Claude Code Agents - Anthropic's agent framework within Claude Code.
• OpenAI Agents SDK - OpenAI's SDK for building multi-agent workflows.
• Anthropic Agent SDK - Build agents with Claude using the Anthropic SDK.
• Pydantic AI - Agent framework built on Pydantic for type-safe agent development.
• Mastra - TypeScript AI agent framework with built-in memory and tools.
• Semantic Kernel - Microsoft's SDK for integrating LLMs into applications.

Agent Infrastructure

Deployment, monitoring, and operational tooling for agent systems.

• The Hive Doctrine — Agent Onboarding Checklist - First 7 days checklist (free).
• The Hive Doctrine — Agent Cost Calculator - Estimate monthly LLM spend (free).
• LangSmith - Observability and evaluation platform for LLM applications.
• Weights & Biases Weave - Tracing, evaluation, and monitoring for LLM apps.
• Helicone - Open-source LLM observability platform.
• Portkey - AI gateway with unified API, caching, and reliability features.
• LiteLLM - Unified interface for 100+ LLM APIs with load balancing.

---

Contributing

Contributions welcome. Please read the contribution guidelines before submitting a PR.

License

Curated by Melisia Archimedes. List licensed under CC0.