skills.rest
jnlanahan
jnlanahanCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

add-ai

Integrate Anthropic LLM capabilities with model selection, prompt caching, and LangSmith tracing.

Updated Apr 30, 2026
Search Tags:#rag#tool use#anthropic#model selection#llm integration#prompt caching#langsmith tracing
One-click install
npx skills add https://github.com/jnlanahan/Product-Agents-and-Skills --skill add-ai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: add-ai
Source: https://github.com/jnlanahan/Product-Agents-and-Skills/tree/main/skills/4-Build-AI
Command: npx skills add https://github.com/jnlanahan/Product-Agents-and-Skills --skill add-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It solves the chaos of adding LLM features to an existing project by standardizing model selection, Anthropic SDK wiring, prompt caching, and observability so integrations are reliable, cost-aware, and testable.

Core Features & Use Cases

  • Model selection by use case: Chooses the right Claude model tier (latency/cost vs deep reasoning) via model-selector instead of guessing.
  • Production-grade Anthropic SDK integration: Sets up the Anthropic client for chat, generation, tool use, structured outputs, and streaming patterns.
  • Prompt caching by default: Enables prompt caching early to reduce latency and cost for repeated system prompts.
  • LangSmith tracing + eval scaffolding: Adds tracing for runs and supports creating datasets + golden-example evals for regression testing.
  • Use cases: Conversational chat, single-turn generation, RAG over knowledge bases, document analysis, image analysis orchestration, and agentic tool-enabled workflows.

Quick Start

Use add-ai when you want to add AI / LLM capabilities to your project, including Anthropic SDK wiring, model selection, prompt caching, and LangSmith tracing and evals.

Frequently Asked Questions about add-ai

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

FAQPage Schema
How do I integrate Anthropic LLMs into an existing software project?▼

Integrate Anthropic LLMs by wiring the SDK, selecting the correct Claude model tier, enabling prompt caching, and validating observability via LangSmith tracing for reliable, cost-aware AI features.

What is the best way to set up prompt caching for Claude models?▼

Set up prompt caching by configuring the Anthropic SDK to cache repeated system prompts early in the integration process, reducing latency and operational costs for repeated generation requests.

How do I add LangSmith tracing and golden-eval datasets to an LLM application?▼

Add LangSmith tracing by instrumenting the Anthropic client to log runs, then create datasets with golden-example evaluations for regression testing to validate observability and ensure consistent model performance.

Does this approach support building RAG and tool use agent loops?▼

Yes, this approach supports building RAG over knowledge bases and agentic tool-enabled workflows by orchestrating document analysis, image analysis, and tool use loops through standardized Anthropic SDK integration.

How do I choose the right Claude model tier for my use case?▼

Choose the right Claude model tier by running project-state detection and using model-selector output to match your latency, cost, and deep reasoning requirements to the appropriate model for chat, RAG, or extraction.

Do I need to hardcode the ANTHROPIC_API_KEY when wiring the SDK?▼

No, you do not need to hardcode the ANTHROPIC_API_KEY when wiring the SDK, as the integration requires configuring the API key securely without hardcoding to maintain production-grade security standards.

Related Skills

Explore More Software Engineering Skills→
raphaelchristiraphaelchristi
43

harness:setup

Configure Harness Evolver, LangSmith, and baseline evaluations for Python projects.

Community
Advanced
tankpkgtankpkg
1

@tank/llm-observability

Diagnose LLM observability and evaluation gaps across production tracing and prompt workflows.

Official
Advanced
vecearvecear

cost-latency-optimizer

Reduce LLM costs and latency via caching, model selection, batching, and prompt optimization.

Community
Advanced
akillnessakillness
40

langsmith

Trace, evaluate, and manage prompts for LLM applications via LangSmith SDKs.

Community
Intermediate