v1.2.0

What's New in v1.2.0

Features

• LLM-based intent classification for REPL input routing
• Session observability pipeline with comprehensive metrics
• Chunked plan generation with outline and wave expansion
• Structured verification report generation
• Deep project analysis and environment preflight checks
• Plan checker with revision loop and coverage gates
• Auto-compression for context window overflow in TUI
• Visual screen transitions (slide-left/right, cross-fade)
• Incremental render diff engine using ANSI cursor positioning
• Flex box layout model with row/column solver, card renderer, modal overlay
• Component system with fuzzy search, focus ring, virtual viewport
• High-contrast accessibility theme and screen reader helpers
• Tool enhancements: Edit 7-strategy cascade, Grep context lines, Glob file type filtering, FileRead partial reads, dangerous command blocklist, structured ToolError type
• Subagent depth limit (MaxAgentDepth=2) and token budget enforcement

Fixes

• Drain stale responses from channels on dispatcher stop
• Short-circuit fallback search on first live provider
• Use staleTTL for cache expiry check
• Recursively skip empty SSE events
• Fix /flush to use tea.ClearScreen
• DNS cache for WebSearch to prevent TOCTOU rebinding (SEC-02)

Performance

• Optimize import graph and symbol indexing in codeintel

Refactoring

• Unify file locking abstraction for cross-platform support
• Consolidate subagent budget exhaustion messages
• Extract forwardMsgToScreen helper
• Replace WriteString with fmt.Fprintf across workflow and metrics

Docs

• Update wiki submodule with Mermaid diagrams and new pages
• Update workflow, features, and docs for major enhancements

v1.1.0

Changelog

• c4da7a3 style(tui): fix formatting and spacing issues in UI components
• 8a98f48 refactor(tui): improve UI rendering and code clarity
• ce1d6ea chore: add reference images 11-14
• aabc0b9 feat(tui): add verification summary header with pass/fail/pending counts
• 67acaad feat(tui): filter out non-chat models from model selectors
• 371a998 feat(tui): add provider-specific API key placeholders and NVIDIA key validation in first-run
• 0559d6b feat(tui): add ETA and animated progress bar to execute phase
• abaaa65 feat(tui): pass tool call description to sidebar on tool start
• 5b771a8 feat(tui): add human-readable tool call descriptions and compact agent badge in sidebar
• 26b617c refactor(tui): replace renderLoading with renderPlaceholder for nil model states
• 37725b3 feat(tui): make toast width responsive to terminal size
• d59abcc feat(tui): add renderPlaceholder helper for instant-loading screens
• 8fa8406 feat(tui): register /complexity command in command registry
• 47e39a9 feat(tui): add complexity slash command handler
• 2560043 feat(tools): register CodeComplexity in default dispatcher
• b63fe2e feat(tools): add CodeComplexity tool for codebase complexity analysis
• d7019c7 feat(errors): extract provider error messages for user-friendly display
• 6a32bc7 fix(provider): detect error objects in SSE stream from NVIDIA/OpenAI-compatible providers
• 8a1e202 feat(nvidia): skip completion-only models from chat UI
• 89841ba feat(provider): detect completion-only models and set Chat flag in capabilities
• ed70fc6 feat(types): add Chat capability flag to CapFlags
• 0efc082 chore: add space.txt to gitignore
• 47f2325 style(tui): fix code formatting and import order in UI components
• 969070b test: add workflow engine tests for ledger integration and context propagation
• 1dfea46 test: update plan_test.go callers for new context-parameter signatures
• b143347 test: update execute_test.go callers for new context-parameter signatures
• 80b115d test: update engine_test.go callers for new context-parameter signatures
• 9c0d176 test: update engine_extra_test.go callers for new context-parameter signatures
• 9256b11 feat: add Path() and Clear() methods to Ledger for factory reset support
• 0ea968c refactor: use shared ledger instance from engine instead of hardcoded path
• 43ddb57 refactor: pass context to buildPlanContext and rename local ctx to planCtx
• 9d5fd1a refactor: pass context to buildExecuteContext and getCodeIntel for cancellation propagation
• 1d8d04f feat: add ledger field to Engine and propagate caller context to codeintel build
• 421e370 test: add sidebar model tests for agent progress and TTL file watcher
• 60025a9 feat: add agent progress tracking, TTL-based file watcher, and computed progress bar
• 29064e5 test: add resume model tests for total count and truncation display
• da85732 feat: show total session count in resume screen when list is capped
• 6cd47fa feat: add totalCount field to ResumeModel for truncation indicator
• fdf68a9 test: add search command tests covering segment content and filtering
• dd23da1 test: update handleReset test for new Cmd-based factory reset flow
• 585b0ed feat: implement /reset as full factory reset and fix /search to include segments
• a01188e feat: add Keychain to CommandContext and update /reset description
• b36ee14 feat: switch sidebar to todo mode on agent loop start and pass keychain to commands
• 0aa76c4 feat: wire sidebar agent progress, subagent lifecycle, and reset flow
• 11f21f5 feat: export ResetCompleteMsg type alias in tui package
• 4cc16cd feat: add ResetCompleteMsg type for factory reset flow
• 80adcbe feat: pass shared ledger instance to workflow engine
• 52af595 feat: register NVIDIA provider at startup and cleanup on force-exit
• 1445469 chore: remove MEDIUM.md documentation file
• 0381d21 chore: remove promotional video file
• 43b9034 docs: add sidebar UI screenshot (image_10)
• 93ef453 docs: add sidebar UI screenshot (image_9)
• bf4421e docs: add sidebar UI screenshot (image_8)
• 787c5b0 chore: add .m31a-worktrees/ to gitignore
• cd6989a test(tui): update settings tests for 3-provider list with NVIDIA
• 57bd920 feat(tui): add NVIDIA short name abbreviation for status display
• 192d95c feat(tui): add NVIDIA NIM to settings UI (provider choice, API key, health)
• 0e3939a feat(tui): add NVIDIA re-registration on settings save
• 605f94e feat(tui): wire NVIDIA API key copy in first-run completion handler
• ef35aed feat(tui): add NVIDIA NIM to first-run wizard provider catalog
• 8af1f9d feat(tui): register NVIDIA NIM in provider registration switch
• 40c766c feat(provider): add NVIDIA NIM provider client
• 8bad3b4 feat(config): add NVIDIA API key resolution and keychain persistence
• ff33aee feat(config): add Nvidia credential and base URL fields to ProviderConfig
• d9a5039 feat(types): add DefaultNvidiaBaseURL constant for NVIDIA NIM API
• 20e4052 feat(images): add additional image assets
• 4a4301b fix(tui): fix toast positioning by returning raw stack
• 3dcd3c9 feat(tui/theme): add tool label styles for all tool types
• 4f142d2 feat(tui): add GetStatus method to subagents model
• 0958ae3 fix(tui/repl): pass max width to logo rendering in welcome screen
• 32775c4 perf(tui/repl): pre-compute wave separator brand style
• dfa2f7a fix(tui/repl): invalidate ThinkingBlock cache on segment close and errors
• 188dfd1 fix(tui/repl): fix blank line accumulation and optimize streaming render
• ce8f42d perf(tui/repl): add ThinkingBlock cache for streaming render
• b7f7115 fix(tui/repl): re-render messages during streaming for animations
• 387186c fix(tui): ensure Provider field is set in phase model picker
• 94d3db0 fix(tui): fix model selector detail pane width overflow
• b306238 fix(tui): ensure Provider field is set on loaded models in selector
• 320b8e9 feat(tui): set max tool calls limit on sidebar initialization
• 6bc0905 fix(tui): fix first run layout overflow on narrow terminals
• 4344e6c fix(tui): fix config tab bar width calculation
• 01270e8 fix(tui/components): fix permission modal layout at narrow widths
• 737450e fix(tui/components): prevent metric card overflow at narrow widths
• 3815b73 fix(tui/components): enable visibility flag on thinking blocks
• ab12141 fix(tui/components): add max width parameter to logo and glow rendering
• b99561b fix(tui/components): fix card gradient line overflow at narrow widths
• d518ea8 feat(tui): add screen name routing and sidebar metric updates in View
• cb154f1 fix(tui): auto-revert sidebar to file tree on phase completion
• 2211c30 feat(tui): integrate sidebar updates into main Update loop
• 99df765 feat(tui): wire TodoWrite callback and fix phase message drain
• f485b76 feat(tui/sidebar): add todo mode, phase pipeline, and metrics tracking
• 1d6729a feat(tui): export new sidebar message types in tui package
• 20fbad1...

M31 Autonomous v1.0.0

M31 Autonomous v1.0.0

Terminal-native AI coding agent that ships, not just suggests.

Installation

Linux (.deb)

sudo dpkg -i m31a_1.0.0_linux_amd64.deb

Linux (.rpm)

sudo rpm -i m31a_1.0.0_linux_amd64.rpm

Linux (.apk)

sudo apk add m31a_1.0.0_linux_amd64.apk

Linux (Arch Linux)

sudo pacman -U m31a_1.0.0_linux_amd64.pkg.tar.zst

macOS / Windows / Linux (tar.gz)
Download the appropriate archive from the assets below.

Checksums

SHA256 checksums are available in checksums.txt