
7/29/2026
What this post added
Introduced 'Plan Mode' for candidates to plan solutions with AI before coding, enhancing visibility into their thought process and tradeoffs. Updated 'AI Fluency' scoring to directly analyze IDE activity and AI interactions, moving beyond chat logs. Added 'Diff View' to compare candidate code with AI contributions, mimicking pull request reviews. Launched 'Plan-Build-Review' repository assessments with pre-built codebases and support for multiple languages and frameworks. Enhanced Chakra AI interviewer with an in-line code editor and deeper technical probing. Bolstered integrity controls with Desktop App enforcement, enhanced proctoring, and language-aware plagiarism thresholds. Introduced a read-only Model Context Protocol server for direct AI model querying of hiring data.