Changelog - JoelBondoux/AtlasMind GitHub Wiki

Changelog

This page highlights major releases. For the complete changelog, see CHANGELOG.md in the repository.


v0.30.4 β€” CI Fixes And Wiki Refresh

  • Fixed the lint issues that were failing CI and restored a passing coverage gate for the currently tested service-layer modules
  • Clarified model-routing documentation around seed models, runtime catalog refresh, and metadata enrichment
  • Added a funding and sponsorship wiki page and refreshed the wiki comparison content

v0.30.3 β€” Copilot Chat Recommendation Restored

  • Restored GitHub Copilot Chat in extension recommendations for the repo and bootstrap templates
  • Updated setup guidance and Copilot runtime wording to point users back to GitHub Copilot Chat

v0.30.2 β€” Copilot Dependency Cleanup

  • Removed the deprecated GitHub Copilot Chat recommendation from the repo and bootstrap templates
  • Updated setup guidance to point to the GitHub Copilot extension instead
  • Renamed Copilot UI/error wording from Copilot Chat to Copilot language model / Copilot Model

v0.30.1 β€” Trust & Freshness Fixes

  • Real daily budget enforcement β€” dailyCostLimitUsd now blocks new requests once the cap is reached
  • Live provider health refresh β€” Status bar updates immediately after key save and model refresh
  • Run Center disk hydration β€” Project Run Center and project runs tree now consume async disk-backed history
  • Settings quick actions β€” Direct buttons for Chat, Model Providers, Project Run Center, Voice, and Vision
  • Budget control in Settings β€” dailyCostLimitUsd is now editable in the Settings panel

v0.30.0 β€” UX & Feature Overhaul

  • Getting Started walkthrough β€” Four-step guided onboarding for new users
  • API key health check β€” Immediate validation after storing a provider key
  • Collapsible settings panel β€” Grouped, collapsible sections replace the flat wall of options
  • Cost persistence and daily budget β€” Session costs persisted to globalState; dailyCostLimitUsd setting with 80%/100% alerts
  • Streaming for Anthropic + OpenAI β€” Full streamComplete() with SSE parsing and tool-call handling
  • Agent performance tracking β€” Success/failure tracking influences future agent selection
  • Cost estimation in plan preview β€” /project shows estimated $low–$high cost before execution
  • Disk-based run history β€” Individual JSON files replace single-blob globalState storage
  • Diff preview in project report β€” File/status table and "Open Source Control" button in report
  • Multi-workspace folder support β€” Quick-pick when multiple folders are open
  • Per-subtask checkpoint rollback β€” Rollback by task ID instead of last-only
  • Memory tree pagination β€” Incremental loading with "Load more…" instead of hard 200-entry cap
  • Provider health status bar β€” Shows how many providers have valid API keys
  • Expanded task profiler β€” 100+ new keywords for more accurate task classification
  • Integration test suite β€” Full orchestrator β†’ agent β†’ cost β†’ performance lifecycle tests

v0.29.0 β€” Constants, Shared Validation & Zod

v0.28.x β€” Project Import & Stability

  • /import command β€” Scan existing workspaces and auto-populate SSOT memory from manifests, READMEs, configs, and license files
  • TypeScript fixes β€” Added "types": ["node"] to tsconfig for full Node.js global support
  • Documentation overhaul β€” Comprehensive README rewrite with logo, comparison table, and complete feature coverage

v0.27.x β€” Skills Gap Analysis & README

  • 11 new skills β€” code-symbols, rename-symbol, code-action, web-fetch, diff-preview, rollback-checkpoint, test-run, diagnostics, file-move, file-delete, git-branch
  • README overhaul β€” Logo, competitor comparison table, comprehensive feature documentation

v0.26.x β€” MCP Integration

  • MCP client β€” Connect external tool servers via stdio or HTTP transport
  • MCP server registry β€” Persistent server configs with auto-reconnect
  • MCP tools as skills β€” External tools seamlessly appear in the skill registry

v0.25.x β€” Project Planner

  • /project command β€” Decompose goals into DAGs of subtasks
  • TaskScheduler β€” Topological sort into parallel batches
  • Ephemeral agents β€” Role-specific agents for each subtask
  • Project Run History β€” Persistent run records with the Run Center

v0.24.x β€” Skill Security Scanner

  • Static analysis β€” 12 built-in rules for custom skill validation
  • Scanner Rules Manager β€” Configure rules via webview panel
  • Pre-enablement gate β€” Custom skills must pass scanning before use

v0.23.x β€” Voice & Vision

  • Voice Panel β€” TTS and STT via Web Speech API
  • Vision Panel β€” Image picker for multimodal prompts
  • /voice and /vision commands

v0.22.x β€” Tool Webhooks

  • Outbound webhooks β€” Forward tool lifecycle events to external HTTPS endpoints
  • Configurable events β€” tool.started, tool.completed, tool.failed
  • Webhook management panel

v0.21.x β€” Cost Tracking & Budget Control

  • CostTracker β€” Per-session, per-provider cost accumulation
  • Budget modes β€” cheap, balanced, expensive, auto
  • Speed modes β€” fast, balanced, considered, auto
  • /cost command

v0.20.x β€” Multi-Agent Orchestration

  • AgentRegistry β€” Custom agents with roles, prompts, and constraints
  • Agent selection β€” Token overlap scoring for best-fit selection
  • Agent Manager Panel β€” Create and configure agents via webview

Earlier Releases

See CHANGELOG.md for the complete version history.