zorost/AI-Engineering-Lab

agent-loop-safety

Wrap an agent loop with step limits, cost caps, human approval gates, and a full trace. Use whenever building or reviewing any tool-calling agent before it touches real systems.

原文の言語: 英語

agent-ops-handoff

Move an agent from laptop demo to operated system, tracing, cost dashboard, scheduled runs, alerting, and rollback. Use when an agent is about to run unattended or serve real users.

原文の言語: 英語

ai-output-review

Review AI-generated code or text before accepting it, spec diff, verifier run, secret scan, and the AI smell list. Use before merging any agent-produced change.

原文の言語: 英語

cloud-deploy-gate

The pre-deployment gate for managed AI platforms (Azure AI Foundry, Google Vertex AI, AWS Bedrock), evals packed, budget set, guardrails on, owner named. Use before any cloud deployment.

原文の言語: 英語

context-budget-audit

Audit the seven claimants on an LLM call's context window, set a working ceiling, and cut in the right order. Use when prompts grow, agents drift, or token bills surprise you.

原文の言語: 英語

error-analysis-50

Read 50 real failures by hand, cluster them into classes, fix the largest class, and extend the golden set. Use whenever an AI system's score stalls or its failures are 'mysterious'.

原文の言語: 英語

eval-first-development

Build the golden set and the automated scorer before touching the prompt, model, or pipeline. Use whenever an AI output's quality will need to be measured, extraction, RAG, agents, classification.

原文の言語: 英語

fine-tune-readiness

Decide whether fine-tuning is justified versus prompting or RAG, and gate the training dataset before any LoRA/SFT/DPO run. Use when someone says 'let's fine-tune'.

原文の言語: 英語

local-model-fit

Compute the VRAM/RAM budget and pick a model size and quantization before downloading anything. Use when choosing local models, planning GPU hardware, or hitting out-of-memory errors.

原文の言語: 英語

mcp-server-craft

Design and build an MCP server whose tools are narrow, typed, idempotent, and documented. Use when exposing any system to AI agents via the Model Context Protocol.

原文の言語: 英語

prompt-suite-versioning

Improve a prompt as a versioned artifact with a score, change one variable at a time, keep the diff, read the failures. Use whenever editing prompts that must stay measurably good.

原文の言語: 英語

rag-pipeline-checkup

Verify a RAG pipeline end-to-end, chunking, embeddings, retrieval quality, reranking, grounded answers with citations. Use when building or debugging retrieval-augmented generation.

原文の言語: 英語

spec-first-ai-feature

Write the one-page spec for an AI feature, user, golden set, metric, gate, refused tradeoffs, before any code or prompt work. Use when starting any AI feature, agent, or pipeline.

原文の言語: 英語

using-zorost-skills

How to pick, run, and amend Zorost agent skills. Use at the start of any task when this catalog is installed, or when a skill seems not to fit.

原文の言語: 英語