feat: B2 Atlas Prompt OS TUI — v0.6.0 MVP#8
Merged
Merged
Conversation
- terminal/bridges/brain-bridge.sh — thin bridge: brain/note/sessions/ decisions/reviews/learn/memory/vault subcommands open vault views. Read-only — never writes. Uses Obsidian URI or open fallback. - mqlaunch.sh — sources brain-bridge alongside hal-bridge - mqlaunch-command-mode.sh — routes brain and shortcut aliases (note, sessions, decisions, reviews) via dispatch_cli_command Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
mqlaunch learn now routes to mq_brain_run "learn" via both dispatch_cli_command and run_arg_command, matching the pattern for note/sessions/decisions/reviews. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…urning to main menu Implements the documented back-marker-pattern so mqlaunch restarts gitlaunch on unexpected exits instead of falling through to the main menu. - gitlaunch: add BACK_MARKER env var, mark_gitlaunch_back() function, call it on b/B before break - mqlaunch: wrap open_git_menu invocation in restart loop; only breaks when back-marker file is written (deliberate b navigation) or after 5 consecutive unexpected exits - smoke test: assert mark_gitlaunch_back, BACK_MARKER, MQ_GITLAUNCH_BACK_MARKER are present in gitlaunch.sh Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- brain-bridge.sh: add `systems` and `verified` subcommands; remove `memory` case arm (memory/ retired in P4) - mqlaunch.sh and mqlaunch-command-mode.sh: extend brain pattern match to include `verified|systems` - tests/brain-bridge-smoke.sh: new smoke test (6 assertions) covering subcommand presence and memory removal Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…(P8)
mqlaunch release-check --brain appends a BRAIN RECORD section after the
standard release-check: runs mq-agent signal --brain . to write the
repo-signal result to mqobsidian/reviews/ via mq-mcp.
mqlaunch-command-mode.sh now passes all args (${@:2}) to
mq-release-check.sh so flags like --brain reach the script.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Three write commands added to command-mode and dispatch table: mqlaunch review-brain [path] → mq-agent review repo --brain mqlaunch signal-brain [path] → mq-agent signal --brain mqlaunch learn-promote <slug> → mq-agent learn promote --approve All three guard on _run_agent availability (declare -f pattern). Covers the two main vault write flows and the human curation gate. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…index Both the panel (show_command_index) and plain-text (show_help) sections updated. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ption 6 - Option 6: was duplicate "signal ." → now "signal --brain ." (saves to brain) - New section SECOND BRAIN: - 15. Review repo → brain (review repo . --brain) - 16. Promote learn pattern (prompts for slug → learn promote --approve) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
_brain_pick_and_promote() lists learn/ files via fzf (falls back to numbered list + read if fzf not found). User no longer needs to know slug from memory. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…inue After pr_aware_push switched branches with git switch -c, the WORK_DIR variable became stale. On next detect_repo iteration, cd "$WORK_DIR" || exit triggered an exit, causing the open_git_menu restart loop to give up after 5 restarts and return to main menu. Fix: pause_git_menu is now called from the case statement in the main loop (same pattern as options 1,2,4,8), not from within run_ai_commit. The continue was also removed — natural loop iteration is sufficient. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MCamner
changed the title
Browses all 43 saved prompts from mqobsidian/_prompts/ via fzf with a 40-line preview panel. Selected prompt is copied to clipboard. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
New b2tui.py CLI with subcommands: - projects: list all 43 prompts by category - run <id>: preview + compose + copy to clipboard + save history - route <task>: keyword-based route matching across 7 routes - validate: check all prompt files readable + route primaries present - history: show recent runs from ~/.b2tui_history.jsonl Wired as mqlaunch b2tui / mqlaunch b2 via both run_arg_command and dispatch_cli_command. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
mqlaunch stack → opens mq-stack/05_RELEASE_STATUS.md in Obsidian mqlaunch roadmap → opens mq-stack/01_ROADMAP.md in Obsidian Routes via mq_brain_run (brain-bridge.sh) consistent with existing brain/note/sessions/decisions entrypoints. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…date order - mq-mcp-review.py: replace direct from-imports with importlib.import_module so review_engine modules resolve via runtime sys.path injection without linter errors; openai unavailability now returns a clear error string - b2tui.py: move errors>0 exit check after route-map warnings so all WARN lines print before early return - ROADMAP.md: add B2 TUI MVP roadmap (phases 0–13); fix list markers to - Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Migrates b2tui from single-file tools/scripts/b2tui.py to a proper
Python package under mqlaunch/b2_tui/ matching the roadmap structure.
- config, models, core/{project_loader,router,prompt_composer,validator,history}
- adapters/obsidian_writer — saves compose runs to ~/mqobsidian/mq-stack/runs/
- CLI: list, categories, show, compose, route, validate, config, history, export-last, open-last
- project_loader reads from PROJECT_INDEX.md as canonical source
- 20 passing tests
- mqlaunch.sh routing updated to new package
- .gitignore: add __pycache__ / *.pyc rules
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
… 21 tests - route() returns (primary, support_routes, matched_keywords) - ROUTE_SUPPORT map drives support project selection - cmd_route prints Primary / Support / Reason blocks - 21 router tests (was 7), 33 total passing Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- compose: empty task rejected immediately with clear error - history: route and compose both saved with command field - history last: show most recent entry - history export: write b2-history.md to mq-stack/runs/ - router: saves route results to history - roadmap: phases 6, 7, 8 marked done Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
dispatch_cli_command() in mqlaunch-command-mode.sh was still calling tools/scripts/b2tui.py directly, shadowing the updated routing in mqlaunch.sh. Updated to PYTHONPATH+module invocation. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- B2App: two-panel layout (categories | projects) + preview pane - j/k navigation, Tab to switch panels, / to search - r=route, c=compose, h=history, v=validate, q=quit bindings - RouteScreen, ComposeScreen, HistoryScreen, SearchScreen modal overlays - mq b2 (no subcommand) now launches the TUI instead of printing help Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…GELOG Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Promotes [Unreleased] to [0.6.0]. All Phase 13 release checks pass: 40 tests green, all CLI commands smoke-tested, docs complete. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
MCamner
changed the title
…ridge All three conflicts were additive. Kept feat/brain-bridge superset (verified, systems, stack, roadmap, b2tui) and merged memory subcommand from main. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
mqlaunch/b2_tui/) — full structured prompt browser and composer for terminal usemq b2 list,show,compose,route,validate,history,config,export-last,open-lastmq b2) — three-panel layout (categories / projects / preview), j/k navigation,/search, compose/route/history modalsAlso includes brain-bridge commands, fzf prompt picker, gitlaunch protected-push improvements, and markdownlint config from earlier work on this branch.
Test plan
PYTHONPATH=. /opt/homebrew/bin/pytest mqlaunch/b2_tui/tests -v— 40 passedmq b2 validate— 43 prompts OK, 0 errorsmq b2 list,categories,show 02.11,route "...",history— all smoke-checked🤖 Generated with Claude Code