Guides – Tutorials and Background on MCP

Use OpenAI Imagegen safely

Set up OpenAI Imagegen safely: the image skill for coding agents with a keyless default mode and an optional CLI fallback.

09.09.2026

Setting up NuGet Manager for disciplined .NET dependency work

A practical guide to loading the skill, checking versions, handling central package management, and reviewing changes.

09.09.2026

Assess Configured Agent for Claude Code safely

A Configured Agent is a custom Claude Code subagent with its own system prompt, restricted tool access, and an isolated context window.

09.09.2026

Develop AI SDK examples with Develop AI Functions Example

Internal Claude skill inside the Vercel AI SDK repository: consistent naming patterns and templates for new AI SDK example scripts.

09.09.2026

Understand Example Skill and structure your own skills clearly

How Anthropic's official Example Skill explains the file format, folder structure, and frontmatter of Claude Code plugin skills as a template.

09.09.2026

Set up the PyTorch AT_DISPATCH_V2 skill

at-dispatch-v2 is an official PyTorch skill that automatically converts AT_DISPATCH macros in ATen C++ code into the new AT_DISPATCH_V2 format.

09.09.2026

Prepare compliance checks safely

Practical guide for compliance checks: define initiatives, organize privacy questions, escalate risks, and secure legal review.

09.09.2026

Use OpenAI Knowledge with source discipline

How the OpenAI Knowledge skill makes AI coding agents base answers on current official docs instead of guessing outdated facts.

09.09.2026

Using OpenAI Figma Implement Design

OpenAI Figma Implement Design guide for Figma MCP setup, node IDs, design tokens, permissions, production review, and safe code handoff.

09.09.2026

Use Agentic Evaluation Patterns with control

Agentic Evaluation Patterns provides guidance that lets coding agents improve their own output through iterative self-critique and evaluation loops.

09.09.2026

Use CLAUDE.md Improver with Claude Code

Anthropic's skill audits and improves CLAUDE.md files in Claude Code projects, with a built-in approval step before edits.

09.09.2026

Using Journal Entry Prep for month-end close

Journal Entry Prep is Anthropic's official finance skill that structures journal entries for accruals, depreciation, and revenue recognition.

09.09.2026

Using Data Validation safely for analysis

Anthropic's validate-data skill QAs analyses before sharing, checking methodology, calculation errors, bias, and misleading charts.

09.09.2026

Using Competitive Analysis safely for product decisions

Practical guide to Anthropic’s official skill for competitive research, comparison matrices, and strategic interpretation.

09.09.2026

Using Power BI Semantic Modeling safely

Practical guide to GitHub Copilot skill guidance for reviewing Power BI semantic models.

09.09.2026

Set up PyTorch Pull Request Review safely

The pr-review skill from the PyTorch repository systematically evaluates pull requests for code quality, test coverage, security, and backward compatibility.

09.09.2026

Use webapp testing safely

Use Playwright-based webapp testing to verify local user journeys, manage servers safely, and avoid fragile or risky browser tests.

09.09.2026

Use Anthropic SQL Queries safely

Practical guidance for dialect selection, query review, performance, and safe database work.

09.09.2026