Updates

Smarter AI Helper & Clearer Project Maps

This update makes the assistant much better at figuring out exactly what you need right away and cleans up how it shows you the structure of huge projects.

• •Smarter Request Handling: The assistant now instantly understands if you're fixing a bug or adding a feature, allowing it to focus its search more effectively and give you faster, more accurate results.
• •Simplified Project Maps: For large projects, we've introduced a high-level 'Discovery View' that helps the agent find the right files quickly while cutting through technical noise.
• •Reliable Background Tasks: We've polished the way the AI handles background work, ensuring consistent behavior across all tasks and providing clearer feedback during loading.

Cleaner Chat Organization & Automatic Naming

This update redesigns how you manage your chat conversations to keep your workspace clean and introduces smart features to help you stay organized automatically.

• •Quick Chat Menu: We replaced the permanent list of chats on the side with a quick pop-up menu that appears only when you need it. You can now rename, switch, or delete your chats directly from this menu, and even save them as files for later.
• •Automatic Chat Titles: Your new chats are now automatically given helpful names based on your first message, so you never have to worry about naming them yourself.
• •Improved Chat Controls: We added a new, faster way to start a fresh chat and updated our buttons to make the app look cleaner and more consistent.

Smarter Project Mapping and Precise Change Tracking

This major update improves how the AI understands your project structure and tracks specific changes in your code, making analysis faster and summaries more accurate.

• •Better Project Understanding: We upgraded the AI's internal toolkit, allowing it to build a much clearer map of your project's structure for more reliable analysis.
• •Crystal-Clear Change Summaries: The AI can now identify exactly which functions or components were modified, providing you with richer and more precise feedback on every change.
• •Stability Polish: Under-the-hood improvements to ensure the system handles complex tasks with maximum resilience and fewer errors.

Smarter Explanations and Precision Code Finding

We've made two major improvements: the AI now communicates changes in plain English, and it explores your project code with blueprint-like precision.

• •Friendly Updates: The AI now gives you friendly, easy-to-read sentences instead of complicated code jargon to explain exactly what it changed. It's like getting a simple summary instead of the full technical report.
• •Blueprint Code Finding: The AI is much smarter about finding the right code to look at. Instead of guessing based on keywords, it now uses a high-precision map of your project's structure, making its code suggestions much more accurate and relevant.
• •Smoother Recovery: The AI is better at handling situations where a file is missing or overlooked. It now recovers more smoothly, leading to fewer interruptions.

Smarter, Safer Editing Tools

We've improved how the AI identifies where to make changes in your code by using unique, collision-resistant markers. This ensures that the AI's instructions never get confused with your actual code, making the editing process much more reliable.

• •Unique Markers: Updated the codes used for finding and replacing text to be completely unique, preventing accidental overlaps with your source code.
• •Strict Error Checking: The system now automatically rejects old or incorrect formats and tries again if it detects a mistake, ensuring higher precision.
• •Consistent AI Behavior: All internal prompts have been updated to use these new safety standards across all models.
• •Updated Documentation: All project guides and architecture maps have been refreshed to reflect these new production standards.

Stronger Agent, Cleaner Output, and Better Investigation Tools

This release focuses on making the Context Agent more robust by giving it specialized investigation tools and implementing security-focused filters to clean up AI responses.

• •Synchronous Context Updates: The AI now always works with the most stable and up-to-date information about your project.
• •Automatic Self-Correction: Improved the system's ability to check its own work and fix mistakes immediately after making changes.
• •Specialized Investigation Tools: Added new ways for the AI to explore your project, such as looking up where functions are defined or checking for code errors.
• •Clean and Secure Output: Implemented a new filter that strips out hidden technical details and strange formatting, ensuring you only see clean and secure text.
• •UI Improvements: Added collapsible sections for cleaner organization in the sidebar and enhanced how code logs are displayed.
• •Resilient Error Handling: The system now automatically rebuilds its understanding of your project if it detects major errors.

Smarter, Faster Project Searching

We've upgraded how the AI explores your project. It now automatically skips over files and folders that aren't important (like temporary data or hidden system files), making it much faster and more accurate when helping you with code.

• •Automatic Filtering: The AI now respects project skip rules (like .gitignore) to stay focused on actual source code instead of temporary files.
• •Faster Performance: Skipping unnecessary folders like node_modules makes finding the information needed to help you much quicker.
• •Improved Accuracy: We've refined the AI's internal logic to be better at picking only the files that are truly relevant to your current task.
• •Updated Guides: All internal documentation has been updated to reflect these smarter searching features and how they improve the AI's performance.

Skip Plan Confirmation & Priority Files Model

This release introduces a streamlined workflow toggle, replaces the legacy heuristic file selector with a smarter Priority Files model, and adds several quality-of-life improvements across the plan execution and self-correction pipelines.

• •Skip Plan Toggle: Added a persistent "Fast Forward" toggle to the sidebar. When enabled, AI-generated plans are executed automatically without requiring manual confirmation, significantly speeding up autonomous workflows for trusted tasks.
• •Priority Files Model: Replaced the legacy heuristic context selector with a git-based Priority Files model. The system now automatically discovers uncommitted files and passes them as priority context to the AI, ensuring the most relevant working-set files are always included.
• •Plan Labels: AI-generated plan explanations now display a "Generated Plan" badge in the chat interface, making them easily distinguishable from regular AI responses.
• •Improved Error Recovery: Ambiguous match and search block errors during file modification steps are now retried autonomously within the execution loop with clarification context, eliminating the need for separate recursive calls and improving reliability.
• •Diagnostic Stabilization: The self-correction workflow now actively waits for per-file diagnostic stabilization before collecting error status, reducing false positives from stale language server data.
• •File Line Counts: Project structure views now include line counts for each file, giving the AI better awareness of file complexity during context selection.

Core Simplification and Legacy Cleanup

This major refactor eliminates legacy dependency graphing and semantic linking logic in favor of more efficient, agentic investigation and modern IDE APIs.

• •Eliminated Deprecated Systems: Removed legacy dependency graph builder, semantic linker, and heuristic context selector. The system no longer relies on slow, static dependency graph generation or TF-IDF based proximity scoring.
• •Modernized Context Selection: Updated context agent tools to use modern IDE symbol and reference APIs. Introduced new context preparation logic to explicitly list clean, warning, and error files for the planning LLM.
• •Enhanced Execution Intelligence: Introduced robust URI resolution in the Plan Executor to handle shorthand paths. Optimized Plan Executor guards to run only when in correction mode, significantly speeding up normal execution.
• •Improved UI Streaming: Modernized the code generator to stream real-time status updates for better feedback during file modifications.
• •Deterministic AI Responses: Enhanced the AI service to capture model thoughts alongside function calls, ensuring more transparent and reliable AI behavior.

Production-Ready Workflow and Smart UX

This update enforces production-readiness by migrating core code extraction logic to deterministic AI function calling and refining context quality. It also introduces significant UX improvements with a "Smart Flash" visual feedback system.

• •Deterministic Code Extraction: Migrated search/replace extraction logic from regex to robust AI function calling, eliminating parsing errors and improving reliability.
• •Smart Flash Feedback: Introduced a Smart Flash effect for AI code edits that highlights the full width of changed lines, intelligently tracks position changes, automatically removes itself when lines are modified, and prevents opacity stacking.
• •Refined Context Quality: Enhanced diagnostics integration and history summarization for more accurate context during long sessions.
• •Governance & Standards: Standardized conditional formatting across modules and updated governance documentation for autonomous workflows.

Smarter Problem Solving and Faster Updates

We've improved how the AI handles errors and applies changes, making it more reliable and efficient.

• •Self-Correction: The AI now automatically checks its own work and fixes any mistakes it makes during the coding process.
• •Automatic Error Detection: The system now scans for code errors immediately after making changes, allowing it to provide instant feedback and repairs.
• •Faster Code Changes: We've optimized how the AI modifies your files, resulting in quicker and more precise updates.
• •High-Performance File Inspection: New background tools allow the AI to quickly scan your files without slowing down your experience.
• •Better Usage Insights: The sidebar now shows more detailed information about your AI usage and request status.

Optimized Performance for Large Projects

This update makes working with large codebases much faster and more efficient by improving how the AI explores your project.

• •Smart Project Scanning: The AI now explores your codebase more intelligently, focusing on what's relevant instead of trying to process everything at once.
• •Improved Search Tools: We've upgraded the AI's ability to find files and understand your project's structure.
• •Faster Loading: Initial loading times for large repositories have been reduced by up to 90%.

Smarter Language Support and Better Intent Recognition

This update makes the AI more consistent across different file types and better at understanding what you're trying to achieve.

• •Unified Language Support: Improved consistency when working with various coding languages across your entire project.
• •Smarter Choice of Context: The AI is now better at deciding which files are most important for your specific task, whether it's fixing a bug or adding a feature.
• •Improved Connection Finding: The AI is now better at following the links between different parts of your code.
• •Behind-the-Scenes Cleanup: Refactored internal tools for better overall speed and reliability.

Faster Thinking and Better Token Tracking

We've switched to a faster AI model for basic tasks and improved how we track your usage limits.

• •Quick Intent Checks: Added a lightning-fast verification step to better understand your requests before starting work.
• •Precise Token Tracking: Improved the accuracy of how your AI usage is measured and reported.
• •Model Upgrade: Switched to the latest efficient AI models for faster response times.
• •Seamless Restarts: Your chat state is now saved and restored even after restarting your editor.

Precise Reading and Better Project Awareness

The AI can now read specific parts of your files more accurately and understands your project's overall setup better.

• •Project Map Integration: The AI now scans your project settings to understand how everything is connected.
• •Targeted Reading: Enhanced the ability to fetch specific lines or sections of code for more relevant investigation.
• •Safer Command Running: Improved the security of internal tools that help the AI understand your project.

Active Project Exploration

The AI can now actively explore your project to find exactly what it needs, within a secure and monitored environment.

• •Secure Investigation Tools: Added a safe way for the AI to search your files and folders using a restricted set of commands.
• •Multi-Step Troubleshooting: The AI can now perform several rounds of exploration to solve complex questions.
• •Action Visibility: You can now see exactly which commands the AI ran directly within the chat window.

Next-Generation AI Support

Updated our systems to support the latest preview versions of advanced AI models.

Instant Action Stopping and Better Stability

Improved the ability to cancel AI tasks immediately and strengthened overall stability during large updates.

Streamlined Task Progress

Simplified how the AI reports its progress, making it easier to follow along with complex changes.

Cleaner Project Scanning

Improved AI focus by automatically ignoring non-text files and refining how it understands your instructions.

• •Binary File Filtering: The AI now skips images and other non-code files to stay focused on what matters.
• •Resilient Planning: Added smarter retry logic to ensure your plans are completed even if there's a minor hiccup.

Automated Documentation Creation

Introduced a new way to automatically generate clear documentation for your project.

Custom Scanning Controls

Added settings to let you control how deeply the AI scans your project for context.

Deeper Structure Analysis

The AI now has a better understanding of how your code is organized and where your cursor is currently located.

Intelligent File Connections

Added a new way for the AI to understand the relationships between different files based on their contents.

Visual Task Timeline

Added a real-time visual timeline to the sidebar so you can track exactly what the AI is doing.

Faster Updates and Safer Undos

Optimized how multiple changes are applied at once and improved the ability to revert changes cleanly.

Better Behind-the-Scenes Management

Improved the internal system that manages how and when changes are applied to your project.

Project Names and Key Management

The sidebar now shows your project's name, and you can manage multiple AI connection keys at once.

Enhanced Security Checks

Strengthened security for all tasks that involve running system commands to ensure your safety.

Visual Usage Statistics

Added a new panel to show you exactly how much of your AI limit you've used.

Improved Multi-Tasking

Updates to how the system handles several AI tasks at once, leading to a smoother experience.

Structured Action Planning

The AI now uses a more reliable method to plan out exactly which steps it needs to take to finish your request.

Better Project Link Awareness

The AI is now much better at understanding how different files and components in your project depend on each other.

Editable Chat History

You can now edit any past chat message to have the AI regenerate its response based on your correction.

More Reliable Action Plans

Improved the AI's ability to create and follow multi-step plans without getting stuck.

Image Upload Support

You can now upload images and screenshots directly to the chat for the AI to analyze.

One-Click Undo and Revert

Added a comprehensive way to undo any changes the AI makes, letting you revert back easily.

Turn Chat into Plans

You can now automatically convert a chat discussion into a structured plan of action.

Intelligent Context Selection

Our file search system is now much smarter at finding exactly which files are needed to answer your questions.

Live Code Writing

Watch the AI write code directly in your editor in real-time.

Privacy-First Architecture

We've transitioned to a local-first design that prioritizes your privacy and project security.

Better Commit Messages and Privacy Updates

Improved the clarity of AI-generated messages and updated how we handle your data for better privacy.

• •Clearer Messages: Refined how AI-generated commit messages are written.
• •Privacy First: Stopped automatic creation of certain database entries for better data control.

Privacy-Safe Error Messages

Ensures that error messages don't reveal sensitive file paths from your computer.

• •Path Sanitization: Replaced full computer paths with simpler names in errors.
• •Internal Updates: Refined internal systems to respect these new privacy rules.

Better Error Recovery and Reliability

Enhanced how the system handles temporary network issues and protects your files from being overwritten by error messages.

• •Automatic Retries: The system now automatically waits and tries again if the AI service is temporarily unavailable.
• •Reliable Connection: Centralized all connections for more consistent behavior.
• •File Protection: Added safeguards to prevent error messages from accidentally being saved into your code files.

Faster Task Cancellation

Improved how quickly tasks stop after you hit the cancel button.

Clearer Help Guides and Policies

Completely overhauled our usage guides and privacy policies to be more transparent and easier to understand.

• •Updated Policies: Better explained how we handle your data.
• •New Usage Guide: Documented how to use new features like smart context and automatic conflict resolution.

Automatic Conflict Resolution

The AI can now automatically detect and fix conflicts that happen when merging different versions of code.

Reliable Plan Execution

Improved the robustness of multi-step plans, including automatic retries for minor errors.

Organized Helpful Links

Added a new clean layout for the links section in settings.

Smarter Plan Reading

Improved how the AI understands its own plans and fixed a minor issue with message formatting.

Sign In and Visual Improvements

Introduced a new sign-in feature and added several visual enhancements to the chat interface.

• •New Sign In: Added a sign-in button and panel for easier access.
• •Better Context Awareness: The AI now understands more about your project's overall structure.
• •Smooth UI: Added loading animations and improved how messages are displayed.

Precise Code Updates

Added an advanced feature to make more precise and reliable changes throughout your code.

First Official Release

The first public release of Minovative Mind on the VS Code Marketplace.

Initial Beta Showcase

The first beta release, focusing on basic project awareness and AI chat functionality.

• •Early Access: Core AI chat features and basic codebase investigation.