Releases: threefoldtech/grid-agent
v0.6.1
v0.6.0
What's New in v0.6.0:
This release primarily focuses on Safety & Reliability for the Grid Agent.
Key Features:
-
🛡️ Advanced Safety System:
- Smart Guard (AI): Automatically assesses risk (Low/Medium/High) and requests approval only for risky actions based on your threshold settings.
- Manual Control: Strict mode requiring approval for every tool execution with per-tool Overrides, provides granular control to auto-approve or always-block specific tools.
- Turbo Mode: Fully autonomous execution (use with caution).
-
⚠️ Beta Disclaimer: New acknowledgment workflow for first-time users. -
🔄 Version Mismatch Detection: The app now alerts you if your local tfcmd binary is out of sync with the GUI.
-
🎲 Smart Node Selection: The CLI now randomizes node selection for ZDB, Kubernetes, and VM deployments to better distribute workloads, instead of always picking the first available node.
-
Documentation: Improved system prompts and persona instructions for better agent reliability.
Full Changelog: v0.5.1...v0.6.0
v0.5.1
Full Changelog: v0.5.0...v0.5.1
v0.5.0
What's New in v0.5.0:
This release focused on User Experience & Documentation:
- 📚 In-App Documentation: Added a comprehensive user guide and troubleshooting tips directly within the app (refactored into a dedicated Docs component).
- 🔒 Security UX: Added "Show/Hide" toggles for sensitive fields like API Keys and Mnemonics in the settings, masking them by default.
- ℹ️ Version Display: Added the application version number to the sidebar footer and settings page for easier tracking.
Changelog:
- feat(gui): add show/hide toggle for API key and mnemonic fields by @mik-tf in #5
- feat(gui): add version display to settings page by @mik-tf in #4
- Help Page by @mik-tf in #7
New Contributors
Full Changelog: v0.4.2...v0.5.0
v0.4.2
Full Changelog: v0.4.1...v0.4.2
v0.4.1
Full Changelog: v0.4.0...v0.4.1
v0.4.0
What's Changed
- feat: Major grid-cli refactor with project-name flag, gateway network extension, and agent improvements by @sameh-farouk in #3
Full Changelog: v0.3.0...v0.4.0
v0.3.0
✨ New Features
🎨 Premium UI/UX Redesign
- Glassmorphism: Implemented a modern glass effect for the header and input area, adding depth and distinct layering to the interface.
- Rich Dark Theme: Switched to a deep "Midnight Blue" palette (
#0f172a) for a more immersive and premium feel. - Vibrant Accents: Replaced standard colors with dynamic gradients (Blue → Cyan) for buttons and user messages.
- Micro-interactions: Added smooth animations for message entry, typing indicators, and hover effects.
🧠 Gemini SDK Migration & Enhanced Intelligence
- SDK Upgrade: Migrated to the new
google.golang.org/genaiSDK and Go 1.24 for improved performance and future-proofing. - Self-Correcting JSON: Implemented a retry loop for LLM responses. If the agent acts out (malformed JSON), the system now feeds the error back to the model for self-correction.
- Workflow Visualization: Introduced refined step types for clearer visualization of the agent's thought process.
👤 Persona Management
- Active Persona Control: New UI elements to clearly display the active persona and allow for easy deactivation.
- Terminology Update: Standardized usage of "Persona" across the application (replacing "Profile").
🔄 System Updates
- Update Checker: Added in-app notification system to alert users when a new version of the application is available.
🚀 Improvements
- System Prompt: Updated the core system prompt to align with new capabilities and improve general reasoning.
- Response Parsing: Refactored LLM response parsing to be more robust, especially for complex tool calls.
🐛 Bug Fixes
- URL Fetching: Fixed an issue where the
read_urltool could hang indefinitely by enforcing a custom HTTP client timeout. - Layout Issues: Resolved text wrapping issues within code blocks to prevent layout overflow in chat messages.
v0.2.0
Highlights
This release focuses on agent intelligence improvements and prompt architecture refactoring for better LLM performance and maintainability.
New Features
- Unified Tool Registry: Dynamic tool registration with
ToolDescriptorinterface for consistent tool documentation and execution - Shared Streaming Execution: Common streaming logic across
CommandToolandtfcmdviaCommandExecutor - Enhanced Error Feedback: Tool failures now include command output in LLM feedback, enabling better error diagnosis
Improvements
Agent & LLM
- Stricter JSON Response Format: Enforced JSON-only responses with explicit rules against markdown wrapping
- Reorganized System Prompt: Condensed from ~190 lines to ~90 lines with clear markdown sections
- Tool-Specific Instructions: Moved CLI rules (flag groups, validation, SSH keys) to respective tool definitions
GUI
- Preserve Profiles on Logout: User profiles now persist after logout (only credentials cleared)
- Error Handling in Workflows: Fixed bug where workflow steps disappeared on error; now shows error as additional step
- Answer/Explanation Separation: Added empty line between answer and explanation in responses
Bug Fixes
- Fixed: Workflow steps disappearing when error occurs mid-workflow
- Fixed: LLM not receiving command output on tool failure (only got "exit status 1")
- Fixed: Answer and explanation concatenated without separation
v0.1.0
First public prototype of Grid Agent: an AI‑powered framework and desktop app for interacting with the ThreeFold Grid. Available for Linux, macOS, and Windows.
Highlights
-
Agent Framework
- Reusable Go library for building AI agents on top of the Grid.
- Pluggable LLM providers (starting with Google Gemini).
- A tool system for executing commands with streaming output.
- Structured JSON responses and workflow orchestration.
-
Grid Agent GUI
- Wails desktop app provides an intuitive chat interface to interact with the ThreeFold Grid using natural language.
- Guided onboarding: mnemonic, network (mainnet / testnet / qanet / devnet), Gemini API key.
- Real‑time, streaming command output.
- Dark/light theme support.
- Profile/Persona manager.
- Chat export.
-
Grid CLI (
tfcmd)- A fork from Grid-CLI that adds Schema export for agent/tool integration.