IDE & Assessments

Code Editor

The editor is built on Monaco (the same engine as VS Code) and supports:

• Syntax highlighting for all major languages
• Multi-file editing with tabs
• Find and replace (Ctrl/Cmd + F)
• Go to line (Ctrl/Cmd + G)
• Keyboard shortcuts matching VS Code defaults
• Auto-save — files are saved automatically as you type

File Explorer

The left panel shows the complete file tree of the task's codebase. You can:

• Click files to open them in the editor
• Expand and collapse directories
• See file type icons for quick recognition

Terminal

A full interactive terminal connected to the sandbox environment via WebSocket. Use it to:

• Run the project's test suite
• Execute shell commands
• Install dependencies
• View logs and debug output
• Resize the terminal by dragging the panel border

AI Chat

The AI assistant panel lets you interact with an AI model during the assessment. Key features:

• Streaming responses — See the AI's response as it's generated in real-time
• Tool calls — The AI can read files, write files, and run shell commands in your sandbox
• Context attachment — Reference specific files or code snippets in your messages
• Conversation history — Full chat history is preserved throughout the session

AI Call Limits

• Free plan: 5 AI calls per session (basic model)
• Pro / Pro+: Unlimited AI calls (full model)
• Team assessments: Unlimited AI calls

Diff Viewer

For code review tasks (task-201 through 205), the IDE includes a diff viewer showing the PR's changes. You can:

• View file-by-file diffs in a split or unified view
• Add inline review comments on specific lines
• Submit a review verdict (approve, request changes, or comment)

Observable Fixtures

Some tasks include observable fixtures — simulated production data that provides context for the problem. Available fixture types:

• Logs — Structured application logs showing errors, warnings, and request traces
• Metrics — Performance dashboards with graphs (response times, error rates, throughput)
• Traces — Distributed tracing data showing request flow through services
• Network — HTTP request/response monitoring
• Alerts — Monitoring alerts and incident notifications
• CI Pipeline — Build and test pipeline status
• Runbook — Incident response playbooks

Task Panel

The right-side task panel displays:

• Task description and requirements
• Expected behavior and acceptance criteria
• Hints (unlocked progressively as time passes)
• Expected test results (how many tests should pass)

Timer & Session Management

Each task has a time limit (typically 30 minutes) displayed at the top of the IDE. Key behaviors:

• The timer counts down from the task's time limit
• When time runs out, the session is automatically submitted
• You can submit early at any time via the Submit button
• Sessions can be paused and resumed (within the overall time limit)
• Idle sessions time out after 90 minutes of inactivity

Submitting Your Work

When you're ready to submit:

1. Click the Submit button in the top-right corner
2. A confirmation modal shows your current test pass rate
3. Confirm submission to end the session
4. After submission, you'll be directed to the debrief — 3-5 comprehension questions about your work
5. Your scorecard is generated (typically 30-60 seconds)