HackerRank's July release continues to advance hiring for the agentic era, giving you deeper insight into how candidates collaborate with AI in development environments that mirror how engineers actually work.

This release introduces the HackerRank MCP, enabling AI models to securely access supported HackerRank resources. It also expands our AI-native assessment experience with richer IDE-based AI Fluency evaluation and new IDE capabilities, including Diff View and Plan Mode, giving you greater visibility into how candidates explore problems, plan solutions, and work with AI throughout an assessment. We've also strengthened assessment integrity with improved proctoring accuracy, helping you evaluate candidates with greater confidence.

Chakra, our AI interviewer, continues to evolve with richer reporting, expanded ATS integrations, improved interview configuration, and stronger interview integrity controls, making it easier to evaluate candidates at scale.

Chakra (AI Interviewer)

Interview Report Improvements

Full Session Video Playback

Interview reports now include full session video playback, allowing recruiters to review the candidate's screen, webcam feed, and interview transcript together. This makes it easier to replay the interview, understand candidate responses in context, and review candidate behavior throughout the session.

Report Export

Recruiters can now export Chakra results for all candidates as a CSV file. The export includes overall scores, section-level scores, and candidate summaries, making it easier to analyze performance, compare candidates, and share results.

Must-Have Requirements for Interview Setup

Recruiters and hiring teams can now define must-have requirements when setting up an interview. These requirements are weighted more heavily during evaluation, helping identify candidates who best match the role's required skills and qualifications.

Candidate Email Invite Template

Recruiters can now edit interview invitation email templates directly within Chakra. They can also set an expiration date for each interview invitation to ensure interviews are completed within the required timeframe.

This helps maintain consistent candidate communication while giving hiring teams greater control over interview scheduling.

Workday Integration

Chakra now supports integrations with Workday, allowing you to manage interviews and review reports within your existing ATS workflows.

API Access

Chakra now provides APIs to programmatically retrieve interview reports and candidate data. This makes it easier to integrate Chakra results into applicant tracking systems (ATS), internal dashboards, and other reporting workflows.

Chakra Integrity

Interview Integrity Enforcement

Chakra now detects when a candidate connects an additional monitor, shares their screen, or exits full-screen, and pauses the interview until the required interview conditions are restored.

During this time, Chakra stops responding and recording resumes only after the required conditions are met.

This helps maintain interview integrity by enforcing required permissions throughout the session.

HackerRank Desktop App

Chakra now supports the HackerRank Desktop App, a lightweight application that candidates download before an interview. Once launched, the app creates a secure interview environment by locking the screen, restricting unauthorized applications, and monitoring candidate activity throughout the session. It automatically flags suspicious activity, including anomalies in webcam feeds and code-writing patterns, and surfaces these insights in the candidate report and session replay to streamline review.

Candidate Experience Improvements

Interview Status Indicators

Chakra now displays its current statuses, such as Listening and Thinking, throughout an interview. This gives candidates better visibility into what Chakra is doing during the conversation and helps create a more transparent interview experience.

Screen

HackerRank MCP

HackerRank now supports a read-only Model Context Protocol (MCP) server, enabling AI models to securely connect to HackerRank.

With read-only access, you can query supported HackerRank resources directly from your AI model. This enables you to build AI-powered experiences, such as building custom dashboards and generating hiring insights.

Improved Leakage Detection

Leakage detection now scans a broader range of web sources to identify exposed questions with greater accuracy. This helps you detect more potentially leaked questions and take action to keep your assessments secure.

Improved AI Fluency Evaluation

AI Fluency evaluation now uses IDE activity instead of only AI chat interactions, giving you a more complete view of how candidates use AI during an assessment.

The updated evaluation captures signals from across the assessment, including how candidates interact with AI and how they apply AI-generated guidance as they work. This provides a richer evaluation of AI fluency, helping you understand how candidates collaborate with AI to solve problems and make progress.

Candidates with limited AI interactions no longer receive an AI Fluency grade when there is not enough data to produce a reliable evaluation.

Test Integrity

Improved Proctoring Detection Accuracy

Several proctoring signals have been enhanced to improve detection accuracy and provide more reliable integrity insights during assessments.

• Object Detection: More accurately identifies and flags mobile phones and tablets in the candidate's webcam feed, including when devices are only partially visible in the frame.

• Screenshot Analysis: Detects a broader range of suspicious applications, including newer AI tools, during Proctor Mode assessments.

• Image Analysis: More reliably detects multiple faces, different individuals appearing during a session, and candidate absence across a wider range of testing environments and lighting conditions, such as home setups, campus labs, and low-light conditions.

• Conversation Detection: Recognizes a broader range of communication patterns in the code editor, including regional languages written in English, giving you greater visibility into potential external communication during assessments.

Language-Aware Plagiarism Thresholds

Plagiarism Detection now uses language-specific thresholds that account for the natural differences in code length across programming languages. This improves detection accuracy for concise submissions while delivering more consistent and reliable plagiarism analysis across assessments.

Gaze Detection (AI Add-on) - Limited Availability

Gaze Detection identifies when a candidate repeatedly looks away from the assessment screen and then resumes typing, helping you detect potential use of external resources during an assessment.

This signal appears in the Integrity Summary and provides additional context when reviewing candidate behavior during an assessment. To reduce false positives, Gaze Detection is reported as a medium-severity integrity signal, allowing hiring teams to review the activity and make informed decisions.

This feature is currently available in limited availability. Feature scope, rollout timelines, and availability are subject to change

More Question Types in Proctor Mode and Desktop App (AI Add-on)

Proctor Mode and the HackerRank Desktop App now support Data Science, QA, DevOps, Whiteboard, and Diagram questions. You can now include these question types in proctored assessments while maintaining test integrity.

Platform

IDE Updates

Diff View for Front-end, Back-end, and Full-stack Questions

Diff View lets hiring teams and interviewers compare a candidate's current project with its original state at the start of an assessment or interview. This makes it easier to understand how a solution evolved over time, including changes made with the AI Assistant.

Diff View is now available for front-end, back-end, and full-stack questions in both Screen and Interview.

Improved AI Assistant Experience

The AI Assistant has been enhanced to deliver faster responses, improved stability, and a more transparent experience across assessments and interviews.

• Faster Responses with Clearer Progress: Surfaces progress as soon as work begins, with a visual thinking trail and organized progress stages that let candidates know it is actively processing their request. Related activities, such as clarifying requests and reading files, are grouped together to reduce visual clutter and make progress easier to follow.

  

• Optimized Context Gathering: Sends only the code and files relevant to each request to the AI model, instead of the entire workspace. This enables faster responses and more accurate context retrieval, especially in larger repositories.

• Automatic Recovery from Transient Errors: Automatically retries transient service errors, such as temporary server errors, before they reach the candidate. This keeps the session uninterrupted instead of displaying an error in the middle of a request.

Plan Mode for AI Assistant (AI Add-on)

The unguarded AI Assistant in Screen and Interview now supports Plan Mode, allowing candidates to create and refine an implementation plan before writing code.

Plan Mode encourages candidates to use AI to explore the problem, organize their approach, and think through implementation before they start coding. This gives hiring teams greater visibility into how candidates collaborate with AI and approach complex problems, not just the code they produce.

Language Updates

Rust is now supported in the HackerRank execution environment. Candidates can use Rust to solve coding questions across supported assessments and interview experiences.

Library Improvements

The HackerRank Library continued to expand with the introduction of Plan–Build–Review style assessments based on code repositories, alongside project-based content, framework-agnostic project assessments, language-specific coding coverage, and AI question libraries.

What’s New

• Introduced Plan-Build-Review repository assessments for Interviews, where candidates work in realistic, prebuilt codebases to resolve customer-style support tickets. Candidates can investigate the issue, use AI to plan a fix or enhancement, implement the solution, and explain their approach during review.

  • Launched new Plan-Build-Review repositories:

    • Workflow (Project management application)

    • LinkUp (Professional networking platform)

    • Shipway (Deployment operations platform)

    • Flagship (Feature flag management platform)

  • Added 159 repository tasks across MERN, Spring Boot, Django, Go, and .NET.

• Introduced 159 project questions across .NET, Angular, React Native, React, Selenium, Spring Boot, and PySpark.

• Added 100 coding questions across Python, TypeScript, JavaScript, C#, Problem Solving, and REST API.

• Added 11 framework-agnostic project questions for C++ and Java.

• Added 20 Prompt Engineering questions.

• Added 20 Retrieval-Augmented Generation (RAG) questions.

Content Additions by Job Family and Skill

SkillUp

AI Assistant for SkillUp Certifications

The AI-assisted IDE is now available in SkillUp certifications. Developers can use AI to build and debug solutions while solving certification challenges. This gives organizations greater visibility into how developers use AI to solve problems and complete tasks.

AI Fluency Insights for Certification Results

SkillUp certification results now include AI Fluency insights alongside skill performance results. This helps organizations better understand how developers use and collaborate with AI during certifications.

Community

Orchestrate

HackerRank Orchestrate is a 24-hour hackathon, designed for the agentic era, where participants build an AI agent to solve an open-ended, real-world problem.

Participants can use any AI tools or agents to design and build their solution. Submissions are evaluated based on the solution, its outputs, the participant's AI interaction transcript, and a personalized AI interview tailored to their work.

This provides a more complete evaluation of AI-assisted problem solving by assessing both what participants build and how they collaborate with AI throughout the process.

Orchestrate is open to everyone, from students to experienced engineers. Each edition introduces a new problem statement and dataset, with participants ranked on a public leaderboard based on their overall performance.

Deprecations and Experience Changes

Screen

Legacy Candidate Site

The legacy candidate site is being deprecated. The site has been in maintenance mode for more than six years and no longer receives security updates.

The AI Add-On package includes our most advanced features that help you assess next-gen developers and maintain test integrity in an AI-native world. To enable these features, contact your account manager or email support@hackerrank.com.

Thank you for supporting our mission to change the world to value skills over pedigree.