skills.rest
ehadziabdic
ehadziabdicCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

llm-agent-orchestration

Build and orchestrate tool-using LLM agents with LangGraph, CrewAI, and OpenAI function calling.

Updated Sep 23, 2026
Search Tags:#react#multi-agent#orchestration#langgraph#crewai#function calling#llm agents
One-click install
npx skills add https://github.com/ehadziabdic/WAgents --skill llm-agent-orchestration-ehadziabdic
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-agent-orchestration
Source: https://github.com/ehadziabdic/WAgents/tree/main/opencode/skills/llm-agent-orchestration
Command: npx skills add https://github.com/ehadziabdic/WAgents --skill llm-agent-orchestration-ehadziabdic

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires openai, langgraph, langchain, langchain-openai, crewai, pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve? Building LLM agents that reason, call tools, remember context, and collaborate reliably requires navigating multiple frameworks and failure modes; this Skill provides tested patterns and runnable scripts for the full agent lifecycle. ## Core Features & Use Cases - Agent Architectures: Implement ReAct, Plan-and-Execute, hierarchical, and multi-agent patterns with LangGraph, CrewAI, and raw OpenAI function calling. - Memory & Human-in-the-Loop: Add short-term, long-term, and episodic memory plus approval gates using LangGraph interrupt() and checkpointers. - Runnable Orchestration Scripts: Use scripts/build_agent.py for a ReAct agent CLI and scripts/multi_agent.py for sequential, parallel, and hierarchical multi-agent runs with execution traces. - Use Case: You need a research pipeline where one agent gathers facts, another writes a report, and a reviewer approves before delivery; run multi_agent.py with the hierarchical pattern and a YAML agent config. ## Quick Start Ask the agent to build a tool-using LLM agent with a calculator and file reader, then run scripts/build_agent.py with your OpenAI API key set.

Frequently Asked Questions about llm-agent-orchestration

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a tool-using LLM agent with OpenAI function calling?▼

Define tools as JSON schemas with strict: True and additionalProperties: false, then loop: send messages with tools, execute any returned tool_calls, append results as tool-role messages, and stop when the model returns content. Cap iterations to prevent runaway loops.

LangGraph vs CrewAI vs AutoGen for multi-agent systems?▼

LangGraph offers graph-based state machines with native checkpointing and high production readiness. CrewAI is declarative and role-based with a low learning curve. AutoGen focuses on conversational multi-turn agents. Choose LangGraph for complex cyclic workflows, CrewAI for team-style collaboration.

How do I add human-in-the-loop approval to a LangGraph agent?▼

Call interrupt() inside a node with a payload for the human, compile the graph with a checkpointer such as InMemorySaver, then resume execution with Command(resume=...). A checkpointer is required because it stores the paused state.

Is it safe to let an LLM agent execute Python code?▼

No, subprocess execution and eval with stripped builtins are not sandboxes; attackers escape via __import__, getattr, or object.__subclasses__(). Use a container with no network and read-only rootfs, or a hosted sandbox like E2B or Modal. For arithmetic, use an AST whitelist evaluator instead.

Why does my LLM agent loop forever or pick the wrong tool?▼

Infinite loops come from missing max-iteration limits; set 10-25 steps. Wrong tool selection usually stems from vague tool descriptions, which are the single most impactful factor in agent accuracy. Keep tools under 15-20 per agent and log every invocation.

Related Skills

Explore More Software Engineering Skills→
ramshan00ramshan00

openai-agents-gemini

Configures OpenAI Agents SDK agents to run on Google Gemini models via the OpenAI-compatible API.

Community
Intermediate
ricablericable

langsmith-observability

Trace, evaluate, and monitor LLM applications across LangChain, OpenAI, Anthropic, and LlamaIndex workflows.

Community
Advanced
alo-expalo-exp
6

ai-llm-safety

Design and implement safety controls for LLM-based agent workflows.

Official
Advanced
zeikarzeikar
3

hyper-auto

Chain hyper-plan-loop into hyper-implement-loop for end-to-end code execution.

Community
Intermediate