AI-Assisted Recruitment Coding Assessment Simulator

An authentic simulation of modern corporate AI-assisted recruitment coding assessments. Prepare for technical interviews and corporate hiring tests in an environment equipped with a guarded scaffolding AI assistant, token budgets, multi-language Monaco editor, automated test runner, and a dual-pillar performance scorecard.

🌟 Key Features

1. Guarded AI Assistant with 5-Step Scaffolding Flow

Unlike general-purpose coding assistants that simply write solutions, this assessment assistant strictly enforces the corporate recruitment scaffolding flow:

1. Phase 1: Problem Framing — Candidates must describe the problem in their own words: inputs, expected outputs, and constraints.
2. Phase 2: Data Structures Planning — Candidates explain which data structures and variables they plan to use and why.
3. Phase 3: Algorithmic Approach — Candidates formulate step-by-step logic and recurrence/base-case reasoning before code is unlocked.
4. Phase 4: Starter Code Stub — Generates tailored starter code with an interactive "Insert in Editor" button.
5. Phase 5: Iterative Refinement & Debugging — Refinements require candidates to explain WHAT logic to change and WHY.

2. Strict DO's & DON'Ts Guardrails

• 🚫 DON'T request complete solutions: Prompts like "write the code", "just solve it", or "give complete answer" are intercepted with an explanation of assessment rules.
• 🚫 DON'T ask to "optimize it" or "fix it" without explanation: The assistant prompts candidates to explain the underlying logic first.
• 🚫 No conversational drift or prompt injection bypasses.

3. Realistic Assessment Constraints & Monaco Editor

• Live Countdown Timer: 20-minute exam timer with visual warning indicators.
• Character Limit: Real-time counter enforcing 500 characters per prompt for clear communication.
• Token Budget Meter: Live tracking of token consumption (Tokens used: X / 2000) with visual progress bar.
• Multi-Language Coding Panel: Supports Python 3.10, C (GCC 11.3), C++, JavaScript (Node 18), and Java (OpenJDK 17).
• Monaco Code Editor: Line numbers, syntax highlighting, bracket colorization, and dark/light themes.
• Integrated Test Runner: Built-in test execution for sample and hidden test cases, reporting Passed/Failed status, Expected vs. Actual output, and execution runtime. Also supports Custom Input testing.

4. Dual-Pillar Performance Scorecard (100 Marks Total)

Evaluates candidates across two equal recruitment pillars upon submission:

• Pillar 1: Real Coding Marks (0 - 50 Marks): Directly based on passing automated test cases against standard and edge cases.
• Pillar 2: AI Assistance & Collaboration Marks (0 - 50 Marks):
  • Scaffolding Progression (0 - 20 pts): Advancement through problem framing, data structures, and approach.
  • Prompt Quality & Specificity (0 - 15 pts): Depth, clarity, and precision of inputs/outputs/constraints.
  • Rule Adherence & Integrity (0 - 10 pts): Following DOs & DON'Ts without asking for direct answers.
  • Token & Prompt Efficiency (0 - 5 pts): Effective utilization of the 2,000-token budget.
• One-Click Scorecard Export: Save your complete performance evaluation and hiring verdict as a Markdown (.md) report.

🚀 Quick Start: How to Use

1. Install: Click Install from the VS Code Marketplace.
2. Open the Simulator:
   • Click the AI Assessment Simulator icon in the left Activity Bar.
   • Or click the $(mortar-board) AI Coding Assessment item in the bottom status bar.
   • Or open the Command Palette (Ctrl+Shift+P / Cmd+Shift+P) and select AI Assessment: Start Practice Assessment.
3. Practice: Browse problems from the Question Bank, collaborate with the AI assistant, write code in the editor, and run test cases.
4. Submit: Click Submit Test when finished to receive your comprehensive 100-mark evaluation scorecard!

⚡ AI Integration: Google Gemini 2.5 Flash

This extension is powered by Google Gemini 2.5 Flash for intelligent scaffolding and candidate evaluation:

• Interactive Key Configuration: Click the AI API KEY button in the AI Assistant header (or run AI Assessment: Configure Gemini API Key from the Command Palette) to connect your free API key from Google AI Studio.
• Encrypted Keychain Storage: Your API key is stored securely in your operating system's keychain (SecretStorage) and is never saved in plain-text configuration files.
• Offline Practice Mode: If no API key is configured or you are practicing offline, the simulator runs automatically using the built-in deterministic guardrail engine with full state-machine support.

📄 License

This extension is licensed under the MIT License.