claudeskills.page

Collection · 12 skills

AI Engineering

Build AI features the way production teams do: integrate the API with retries and streaming, write prompts against an eval set, ground answers with retrieval, force structured outputs, design tools for agents, cut costs with caching, and defend against prompt injection.

Best for Engineers adding LLM-powered features to a product.

What's inside

Listed in the order we suggest adopting them. You don't call them by name: each skill loads on its own when a request matches it.

  1. LLM App Builder Ship LLM features with timeouts, streams, and budgets AI Engineering llm-app-builder 225 lines
  2. Prompt Engineering Prompts measured against evals, not eyeballed AI Engineering prompt-engineering 200 lines
  3. LLM Evals Know whether the change helped, with error bars AI Engineering llm-evals 173 lines
  4. RAG Pipelines Fix retrieval first, then make answers cite it AI Engineering rag-pipelines 191 lines
  5. Vector Search ANN tuned against exact search, with filters on AI Engineering vector-search 210 lines
  6. Structured Outputs Schema-valid JSON, then checks on what it says AI Engineering structured-outputs 220 lines
  7. Tool-Use Agents Tools the model can pick right, loops that stop AI Engineering tool-use-agents 209 lines
  8. Prompt Caching and Costs Cut the LLM bill with caching before downgrades AI Engineering prompt-caching-costs 189 lines
  9. Prompt Injection Defense Contain what an injected instruction can do AI Engineering prompt-injection-defense 169 lines
  10. LLM Observability Trace any bad answer back to its exact prompt AI Engineering llm-observability 196 lines
  11. LLM Gateway Routing Fallbacks that are tested, retries that cannot snowball AI Engineering llm-gateway-routing 263 lines
  12. Content Moderation Policy first, thresholds from data, humans for the gray zone AI Engineering content-moderation 224 lines

Keep exploring

Other collections

All collections →
esc