Agent Design Patterns: 510 curated resources
Curated material on agent design patterns within agents and harnesses.
510 resources · page 15 of 16
aider docs: Repository map
Creator not supplied; see original source.
Aider documentation on its repository map, a concise tree-sitter summary of a codebase given to the LLM as context.
Anomalib
Creator not supplied; see original source.
Anomalib documentation: deep-learning library of anomaly detection algorithms with training, benchmarking and deployment tools.
Building a C compiler with a team of parallel Claudes
Creator not supplied; see original source.
Anthropic engineering case study of building a C compiler with many Claude agents working in parallel.
CEPEJ assessment tool for AI in judicial systems
Creator not supplied; see original source.
Council of Europe CEPEJ assessment tool (PDF) operationalising its ethical charter on AI in judicial systems.
A harness for every task: dynamic workflows in Claude Code
Creator not supplied; see original source.
Claude blog post on dynamic workflows in Claude Code, where it writes and orchestrates its own multi-agent harness.
Code Mode: the better way to use MCP (Cloudflare Blog)
Creator not supplied; see original source.
Cloudflare post on Code Mode, where agents write TypeScript against MCP APIs instead of calling tools directly.
Don't Build Multi-Agents (Cognition)
Creator not supplied; see original source.
Cognition post by Walden Yan arguing for shared context and single-threaded agents over naive multi-agent designs.
Multi-Agents: What's Actually Working (Cognition)
Creator not supplied; see original source.
Cognition follow-up post on which multi-agent setups work in production and where single agents still fit.
Effective context engineering for AI agents (Anthropic)
Creator not supplied; see original source.
Anthropic post on context engineering for agents: treating context as finite, with compaction, memory and tool selection.
Engineering fast, performant AI agents through fine-tuning (Decagon)
Creator not supplied; see original source.
Decagon engineering post on using several fine-tuned smaller models to cut latency in customer-service agents.
DIPY
Creator not supplied; see original source.
DIPY documentation: Python library for diffusion MRI and 3D/4D neuroimaging processing.
Direct Preference Optimization
Creator not supplied; see original source.
arXiv paper (2023) introducing Direct Preference Optimization (DPO), preference training without a separate reward model.
DSPy docs: Optimization overview
Creator not supplied; see original source.
DSPy documentation on optimisers such as MIPRO and GEPA for tuning prompts and weights programmatically.
Effective harnesses for long-running agents
Creator not supplied; see original source.
Anthropic post on harnesses for agents that work across many sessions, using an initializer agent and incremental progress.
FINOS AI Governance Framework
Creator not supplied; see original source.
FINOS AI Governance Framework: open-source controls for adopting AI in financial services, mapped to risks.
How we built our multi-agent research system (Anthropic)
Creator not supplied; see original source.
Anthropic post on how it built Claude's multi-agent Research system with an orchestrator and subagents.
Flower documentation
Creator not supplied; see original source.
Flower documentation: open-source framework for federated learning, simulation and deployment.
Google Agent Development Kit
Creator not supplied; see original source.
Google Agent Development Kit (ADK) documentation for building, evaluating and deploying multi-agent systems.
Harness design for long-running application development
Creator not supplied; see original source.
Anthropic post on harness design for agents that build applications over long runs.
How we contain Claude across products
Creator not supplied; see original source.
Anthropic engineering post describing the containment architecture used to isolate Claude across its products.
12-Factor Agents
Principles for agents that hold up in production.
Intercom and Fin developer platform
Creator not supplied; see original source.
Intercom developer platform: APIs, SDKs and integration guides for Intercom and its Fin AI agent.
JAX documentation
Creator not supplied; see original source.
JAX documentation: Python library for accelerator-oriented array computing with automatic differentiation, JIT compilation and sharding.
Kimi K2.5: Visual Agentic Intelligence
Creator not supplied; see original source.
arXiv paper (2026) Kimi K2.5 technical report on visual agentic intelligence.
Kimi K2: Open Agentic Intelligence
Creator not supplied; see original source.
arXiv paper (2025) Kimi K2 technical report: MuonClip optimizer and agentic post-training.
Context Engineering with Deep Agents (LangChain)
Creator not supplied; see original source.
LangChain post on context engineering in Deep Agents using offloading, summarisation, skills, subagents and memory.
LLM Powered Autonomous Agents
Classic overview of planning, memory and tool use in agents.
LlamaIndex blog
Creator not supplied; see original source.
LlamaIndex blog with posts on document agents, retrieval, LlamaParse and agent engineering patterns.
Long Horizon Document Agents (LlamaIndex)
Creator not supplied; see original source.
LlamaIndex post on designing long-horizon agents that process and act on documents.
LoRA
Creator not supplied; see original source.
arXiv paper (2021) "LoRA: Low-Rank Adaptation of Large Language Models", introducing LoRA, low-rank adaptation for fine-tuning large models.
MAPIE
Creator not supplied; see original source.
MAPIE documentation: Python library for conformal prediction intervals and sets for any model.
MCP blog: The 2026-07-28 Specification
Creator not supplied; see original source.
MCP blog post on the 2026-07-28 specification release, summarising protocol changes and the deprecation policy.