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.
- Agent Memory
- Agent Security
- Multi-Agent Governance
- Agent Identity & Personas
- Agent Evaluation
- MCP Resources
- Agent Commerce
- Multi-Agent Frameworks
- Agent Infrastructure
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.
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.
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.
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.
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.
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.
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.
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.
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.
Contributions welcome. Please read the contribution guidelines before submitting a PR.
Curated by Melisia Archimedes. List licensed under CC0.
