All skills
Skillintermediate

Agent Knowledge Library

Index of reference documents for AI agent development, prompt engineering, and platform integration.

Claude Code Knowledge Pack7/10/2026

Overview

Agent Knowledge Library

Index of reference documents for AI agent development, prompt engineering, and platform integration.

General Knowledge

TopicFileKey Content
Agent ArchitectureAI-AGENT-ARCHITECTURE-RESEARCH.mdReasoning patterns (ReAct, CoT, ToT, GoT, Reflexion), state management, agent loops
Prompt EngineeringPROMPT-ENGINEERING-REFERENCE.mdSystem prompts, few-shot, chain-of-thought, structured output, agentic prompting
Tool Use & MCPFUNCTION-CALLING-TOOL-USE-REFERENCE.mdFunction schemas, parallel calls, MCP protocol, tool security
Context OptimizationCONTEXT-OPTIMIZATION-REFERENCE.mdToken budgeting, RAG, prompt caching, lost-in-middle problem
Multi-Agent SystemsMULTI-AGENT-SYSTEMS-REFERENCE.mdOrchestration patterns, frameworks (LangGraph, AutoGen, CrewAI), communication
Instruction FollowingLLM-INSTRUCTION-FOLLOWING-RELIABILITY.mdInstruction hierarchy, guardrails, hallucination prevention, reliability

Platform Integration

PlatformFileKey Content
Claude CodeCLAUDE-CODE-REFERENCE.mdHooks, skills, MCP, Agent SDK, subagents, configuration
OpenCodeOPENCODE-REFERENCE.mdCommands, agents, MCP integration, project instructions
Codex CLICODEX-REFERENCE.mdSkills, sandbox, MCP, session management

Project-Specific

TopicFileKey Content
Workflow Agentsworkflow.md/next-task phases, /ship workflow, agent tool restrictions
Release Processrelease.mdVersion bumping, GitHub Actions, npm publishing

Topic Ownership

To avoid duplication, each topic has one canonical source:

  • Reasoning patterns (ReAct, Plan-and-Execute, ToT) → AI-AGENT-ARCHITECTURE-RESEARCH.md
  • Prompting techniques (CoT, few-shot, structured output) → PROMPT-ENGINEERING-REFERENCE.md
  • Tool/function calling (schemas, MCP) → FUNCTION-CALLING-TOOL-USE-REFERENCE.md
  • Context management (RAG, caching, tokens) → CONTEXT-OPTIMIZATION-REFERENCE.md
  • Multi-agent coordination (orchestration, frameworks) → MULTI-AGENT-SYSTEMS-REFERENCE.md
  • Reliability (guardrails, hallucinations) → LLM-INSTRUCTION-FOLLOWING-RELIABILITY.md