agentleFS
MicrosoftKnown publishermicrosoft-foundry/foundry-agent-webapp/.github/skills/writing-typescript-code/SKILL.md

writing-typescript-code

Skill124 starsChanged 7 months agoMITView on GitHub

Provides TypeScript and React coding standards for this repository. Use when writing or modifying TypeScript code, creating React components, implementing MSAL authentication, or working with the frontend.

  • Reads credentials
  • Sends data out

The file

We split this skill into the rules it gives an agent

12 rules. About 0 are common, 6 are distinctive, and 6 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Goal: Write type-safe React components with proper MSAL integration

    line 8
  2. Distinctive:

    The frontend runs with Vite HMR: When you edit TypeScript/React code:

    line 12
  3. Distinctive:

    Check the terminal: - Look for HMR updates in the "Frontend: React Vite" terminal

    line 15
  4. Distinctive:

    Testing changes: Use Playwright browser tools to:

    line 24
  5. Distinctive:

    Check browser console logs for state transitions and errors

    line 26
  6. Distinctive:

    Use functional components + hooks + typed props:

    line 45
  7. Very specific:

    Use useState (local) or Context API (shared):

    line 95
  8. Very specific:

    Use useMemo and useCallback for expensive computations and stable references:

    line 105
  9. Very specific:

    Extend state: Add discriminated action to AppAction union in frontend/src/types/appState.ts

    line 236
  10. Very specific:

    Handle in reducer: Update frontend/src/reducers/appReducer.ts (keep pure, no side effects)

    line 237
  11. Very specific:

    Create service method: Add to ChatService if network interaction needed

    line 238
  12. Very specific:

    Wire container: Update AgentChat.tsx to dispatch actions

    line 239

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

TypeScript Coding Standards

Goal: Write type-safe React components with proper MSAL integration

Hot Module Replacement (HMR) Workflow

The frontend runs with Vite HMR. When you edit TypeScript/React code:

  1. Save the file - Vite instantly updates the browser (no refresh needed)
  2. Check the terminal - Look for HMR updates in the "Frontend: React Vite" terminal
  3. State is preserved - React state persists through most edits

VS Code Tasks (use Run Task command or check terminal panel):

  • Frontend: React Vite - Runs npm run dev with HMR enabled
  • Logs are visible directly in VS Code terminal

No restart needed - Just edit, save, and see changes instantly in the browser.

Testing changes: Use Playwright browser tools to:

  • Navigate to http://localhost:5173
  • Check browser console logs for state transitions and errors
  • Inspect network requests for API validation

TypeScript Config

Enable strict mode + explicit types (avoid any):

{
  "compilerOptions": {
    "strict": true,
    "noImplicitAny": true,
    "strictNullChecks": true
  }
}

React Components

Use functional components + hooks + typed props:

interface MessageProps {
  message: string;
  sender: 'user' | 'agent';
}

function Message({ message, sender }: MessageProps) {
  return <div className={`msg-${sender}`}>{message}</div>;
}

MSAL Pattern

Always: Try silent first, fallback to popup:

try {
  const { accessToken } = await instance.acquireTokenSilent({
    ...tokenRequest,
    account: accounts[0]
  });
  return accessToken;
} catch {
  const { accessToken } = await instance.acquireTokenPopup(tokenRequest);
  return accessToken;
}

Environment Variables

CRITICAL: Access at module level only (build-time replacement):

// ✅ Correct - module level
const clientId = import.meta.env.VITE_ENTRA_SPA_CLIENT_ID;

// ❌ Wrong - inside function (won't work after build)
function getClientId() {
  return import.meta.env.VITE_ENTRA_SPA_CLIENT_ID;
}

Available variables:

  • VITE_ENTRA_SPA_CLIENT_ID - Entra app client ID
  • VITE_ENTRA_TENANT_ID - Azure tenant ID

State Management

Use useState (local) or Context API (shared):

const [messages, setMessages] = useState<Message[]>([]);
const [loading, setLoading] = useState(false);
const [error, setError] = useState<Error | null>(null);

Memoization Patterns

Use useMemo and useCallback for expensive computations and stable references:

// Memoize computed values
const isAuthenticated = useMemo(
  () => accounts.length > 0,
  [accounts.length]
);

// Memoize callbacks to prevent child re-renders
const getAccessToken = useCallback(async () => {
  // ... token acquisition logic
}, [instance, accounts]);

// Return memoized object for stable reference
return useMemo(
  () => ({ getAccessToken, isAuthenticated, user }),
  [getAccessToken, isAuthenticated, user]
);

API Calls

Include Authorization header + use async/await:

const response = await fetch('/api/endpoint', {
  method: 'POST',
  headers: {
    'Authorization': `Bearer ${token}`,
    'Content-Type': 'application/json'
  },
  body: JSON.stringify(data)
});

if (!response.ok) throw new Error(`API error: ${response.status}`);

npm Dependencies

frontend/.npmrc sets legacy-peer-deps=true automatically — no flag needed when running from frontend/.

Gotcha: legacy-peer-deps skips automatic peer dependency installation. If a package requires peer deps, add them explicitly to package.json.

Example: @lexical/yjs requires yjs as a peer dependency. Since peer deps aren't auto-installed, yjs must be in package.json directly.

Before committing package changes, verify with:

npm ci  # Fails if lock file is out of sync with package.json

Common Mistakes

  • ❌ Accessing import.meta.env.* in functions
  • ❌ Calling hooks conditionally or in loops
  • ❌ Using any type
  • ❌ Storing tokens in component state
  • ❌ Running npm install outside frontend/ (misses .npmrc config)
  • ❌ Missing memoization in custom hooks (causes infinite re-renders)
  • ❌ Returning new objects from hooks without useMemo

Project-Specific: Architecture

ConcernImplementation
State ManagementCentralized Context + useReducer (AppContext) with discriminated action union
AuthenticationMSAL redirect flow; silent token refresh; useAuth hook
Chat StreamingSSE in ChatService with abort controllers for cancellation
AccessibilityLive region (aria-live), aria labels, focus management
LoggingDev-only diff-based logger

Project-Specific: Key Components

ComponentPurpose
AgentChat.tsxContainer wiring chat state to controlled ChatInterface
ChatInterface.tsxStateless controlled UI; renders messages, input, errors, BuiltWithBadge
chat/AssistantMessage.tsxMemoized assistant message with streaming + citation footnotes
chat/UserMessage.tsxMemoized user message with image thumbnail previews
chat/ChatInput.tsxFile uploads, character counter, cancel streaming button
chat/CitationMarker.tsxInline superscript citation badge with tooltip + click handler
core/Markdown.tsxRenders markdown with inline citation markers via ContentWithCitations
core/BuiltWithBadge.tsx"Built with Microsoft Foundry" link badge (centered under input)

Project-Specific: Citation System

Parser: frontend/src/utils/citationParser.ts

Handles Azure AI Agent citation formats:

  • Assistants/Responses API: 【4:0†source】, 【13†myfile.pdf】
  • Azure OpenAI On Your Data: [doc1], [doc2]

Flow:

  1. parseContentWithCitations() replaces placeholders with [N] markers
  2. Markdown.tsx renders CitationMarker components for each [N]
  3. Clicking inline marker scrolls to footnote (with highlight animation) or opens URL
  4. AssistantMessage.tsx renders footnote list with icons by type (URI/file/document)

Key Types (frontend/src/types/chat.ts):

  • IAnnotation - Citation metadata (type, label, url, fileId, quote, textToReplace)
  • IndexedCitation - Parsed citation with display index

Project-Specific: File Upload Validation

Limits: 5MB per file, max 5 files total

See: frontend/src/utils/fileAttachments.ts for validateImageFile() and validateFileCount()

Project-Specific: ChatService

File: frontend/src/services/chatService.ts

Key patterns:

  • Class-based service with Dispatch<AppAction> for state updates
  • AbortController for stream cancellation (cancelStream())
  • retryWithBackoff() for resilient API calls (3 retries, 1s initial delay)
  • SSE parsing via parseSseLine() and splitSseBuffer() utilities
  • Duplicate chunk suppression to prevent UI flicker

Methods:

MethodPurpose
sendMessage()Orchestrates auth, file conversion, streaming
cancelStream()Aborts active stream, dispatches CHAT_CANCEL_STREAM
clearChat()Resets conversation state
clearError()Clears error without affecting chat

Project-Specific: Adding Features

  1. Extend state: Add discriminated action to AppAction union in frontend/src/types/appState.ts
  2. Handle in reducer: Update frontend/src/reducers/appReducer.ts (keep pure, no side effects)
  3. Create service method: Add to ChatService if network interaction needed
  4. Wire container: Update AgentChat.tsx to dispatch actions
  5. Update UI: Pass callbacks to controlled component

Project-Specific: Accessibility Checklist

  • ✅ Live region announces latest assistant message
  • ✅ aria-busy attribute on messages container during streaming
  • ✅ Buttons have aria-label when icon-only
  • ✅ Focus returns to input after sending
  • ✅ Character counter linked via aria-describedby

Related Skills

  • implementing-chat-streaming - SSE streaming patterns and frontend state flow
  • troubleshooting-authentication - MSAL popup issues and token debugging
  • testing-with-playwright - Browser testing and accessibility validation

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.