This gives you a proper agent architecture that decouples your AI logic from whatever interface you bolt onto it. The core agent uses OpenRouter's SDK for model access and emits granular events (stream deltas, tool calls, reasoning updates) through an EventEmitter, so you can drive an Ink terminal UI, HTTP API, or Discord bot from the same codebase. The items-based streaming model is smart: instead of accumulating chunks, you replace items by ID as they update, which keeps state management clean. Includes Zod for tool definitions and handles extended thinking for models that support it. Good default choice if you're building something that might need multiple frontends or you just want your agent logic testable without rendering anything.

skills.sh installs
2.7K
GitHub stars
29
last updated
Jul 14, 2026

Install

$ npx -y skills add openrouterteam/agent-skills --skill create-agent --agent claude-code

Installs into .claude/skills of the current project.

Files

Build a Modular AI Agent with OpenRouter

This skill helps you create a modular AI agent with:

  • Standalone Agent Core - Runs independently, extensible via hooks
  • OpenRouter SDK - Unified access to 300+ language models
  • Optional Ink TUI - Beautiful terminal UI (separate from agent logic)

Architecture

Sponsors

Details

Repository
github.com/openrouterteam/agent-skills
Category
AI & Agent Building

More from openrouterteam/agent-skills

All 2
1steipete avataropenai-whisper-apisteipete/clawdisopenai whisper api2.7Kskills.sh installs
2replicate avatarprompt-videosreplicate/skillsPrompting techniques for AI video generation models on Replicate. Use when writing prompts for video models or building video generation features.2.7Kskills.sh installs
3yyh211 avatardaily-ai-newsyyh211/claude-meta-skillAggregates and summarizes the latest AI news from multiple sources including AI news websites and web search. Provides concise news briefs with direct links to 2.6Kskills.sh installs
4yanhongxi-openclaw avatarproactive-self-improving-agentyanhongxi-openclaw/proactive-self-improving-agentproactive self improving agent2.6Kskills.sh installs
5nvidia avatarrag-evalnvidia/skillsFilesystem RAG benchmarks: corpus/, train.json, evaluate_rag.py (RAGAS quality). Not for prod monitoring, latency/throughput benchmarking (use rag-perf), or eva2.5Kskills.sh installs
6huggingface avatarhuggingface-llm-trainerhuggingface/skillsTrain or fine-tune language and vision models using TRL (Transformer Reinforcement Learning) or Unsloth with Hugging Face Jobs infrastructure. Covers SFT, DPO, 2.4Kskills.sh installs