How to use the C# Async Best Practices agent skill

What does the csharp-async skill make the agent do?

The skill is a single instruction file, SKILL.md. It is not a tool and it does not run anything itself. It changes how the agent thinks about your C# code when you write or review it.

The guidance covers four areas: naming, return types, exception handling and performance, plus common pitfalls and implementation patterns. When you ask the agent to review your code, it is asked to identify issues against these practices and suggest improvements.

• Naming: use the Async suffix on async methods and match names to synchronous counterparts where it applies, such as GetDataAsync() for GetData().
• Return types: return Task<T> or Task, consider ValueTask<T> for high-performance cases, and avoid async void except for event handlers.
• Exceptions: wrap await expressions in try/catch, do not swallow exceptions, and consider ConfigureAwait(false) in library code.
• Performance: use Task.WhenAll() for parallel work, Task.WhenAny() for timeouts, and cancellation tokens for long-running operations.
• Pitfalls: never use .Wait(), .Result or .GetAwaiter().GetResult() in async code, and always await Task-returning methods.

How do you install the csharp-async skill?

Download the ZIP for csharp-async from this page. The skill ships as a folder that contains SKILL.md. Unzip it so the csharp-async folder sits in one of the locations below.

For a personal skill available in all your projects, unzip it into ~/.claude/skills/. To share it with one project, unzip it into that project's .claude/skills/ folder. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

• Personal: ~/.claude/skills/csharp-async/
• Project only: .claude/skills/csharp-async/ inside your project folder
• Claude apps: Settings → Capabilities → Skills, then upload the ZIP

How do you trigger the skill with a natural request?

The skill's description is 'Get best practices for C# async programming'. Ask for C# async help in plain language and mention async, await, or the C# code you want checked. You do not need a special command.

Good requests name the code and the goal. For example, ask the agent to review a specific file, or to check a method for blocking calls. Because the skill's instructions end by asking the agent to identify issues when reviewing your code, reviews are the most direct use.

• Review this C# service class for async problems and suggest fixes.
• Does this method use async correctly? Check return types and naming.
• Find any .Result or .Wait() calls in this code that could deadlock.
• Write an async method that fetches orders and returns a Task<List<Order>>, following C# async best practices.

What are some practical examples?

The skill is most useful where the same async rules need to be applied again and again. Each example below uses the guidance in SKILL.md; the agent's exact output will vary with your code.

A reviewer might paste a method that calls GetData().Result and ask what is wrong. The guidance says never to use .Result in async code, so the agent should point out the blocking call and suggest awaiting the method instead. Another common case is a loop that awaits tasks one at a time: the agent can suggest Task.WhenAll() for parallel execution.

What does the skill need, and is it safe to use?

The skill lists no requirements. It does not run scripts and does not use the network. It is a single SKILL.md file of written guidance, so there is no code from the skill to execute.

Your agent still reads your code and may send it to the model provider as part of normal use, as with any request. The skill itself adds only instructions.

• Requires: nothing listed
• Runs scripts: no
• Uses network: no
• Files: SKILL.md

What tips help you get better results?

Keep the scope small. Ask the agent to review one class or one method at a time, since the guidance is brief and general.

Treat suggestions as a checklist, not a final answer. The skill does not replace testing, profiling, or review of framework-specific behaviour.

• Give the agent the full method, including its await calls, so it can judge the flow.
• Use it for team consistency: share the same skill folder so everyone gets the same review criteria.
• For legacy code, ask specifically about blocking calls such as .Result and .Wait() to find deadlock risks.