Turn scattered shell commands into structured workflows.
π View case study: mcamner.github.io/macos-scripts/case.html
Stop memorizing commands. Start running workflows.
curl -fsSL https://raw.githubusercontent.com/MCamner/macos-scripts/main/install.sh | bashgit clone https://github.com/MCamner/macos-scripts.git
cd macos-scripts
./install.shmqlaunch doctorThis will:
- verify your environment
- check required dependencies
- validate your setup
- highlight issues with clear fixes
mqlaunch- browse workflows via the interactive menu
- or run commands directly (
perf,system,dev,tools)
Instead of:
top
df -h
ps aux | sort -nrk 3 | head
./tools/scripts/system-check.shRun:
mqlaunch perfMost environments don't lack tools β they lack structure.
This project turns:
useful chaos β usable system
One command β structured workflows β repeatable execution
- single entrypoint:
mqlaunch - organized workflows (Dev, System, Performance, Git, Release, Tools)
- works as interactive menu and direct CLI
- built-in AI for questions, code generation, and error fixes
mqlaunch # interactive menu
mqlaunch perf # performance overview
mqlaunch system check # system health
mqlaunch dev # dev workflows
mqlaunch git # git menu
mqlaunch release # release menu
mqlaunch ask "your question" # AI repo assistant
mqlaunch srm ask "your question" # Semantic Repository Memory assistant
mqlaunch fix "error or task" # AI generates copy-paste shell commands
mqlaunch atlas # AI REPL session
mqlaunch doctor # environment check
mqlaunch selftest # smoke tests + shell lint
mqlaunch demo # guided demoAsk questions about the repo directly in the terminal:
mqlaunch ask "how does command routing work?"
mqlaunch ask "what does doctor.sh check?"The repo assistant uses Semantic Repository Memory by default:
MQ_REPO_VECTOR_STORE_ID=vs_your_repo_store_idQuery a specific Semantic Repository Memory store with the Responses API:
mqlaunch srm inspect
mqlaunch srm ask "what repo is indexed here?"
SRM_VECTOR_STORE_ID=vs_your_semantic_memory_store_id mqlaunch srm search "vector store upload flow"Start an interactive session with Atlas β a senior systems engineer embedded in mqlaunch:
mqlaunch atlasatlas> why is my push being blocked?
atlas> explain the release flow
atlas> exit
Atlas answers inline without opening a browser.
Get copy-paste ready shell commands for errors and tasks:
mqlaunch fix "ERROR: CHANGELOG does not contain version 0.1.9"
mqlaunch fix "write a function that checks if a port is in use"
mqlaunch fix "add option 13 'View stash' to the git menu"Output is always runnable β cat heredocs, sed, mkdir β no theory:
# Add changelog section for 0.1.9
cat >> CHANGELOG.md << 'EOF'
## [0.1.9] - 2026-05-07
### Changed
* ...
EOFCopy, paste, run. Uses repo context from the vector store.
Auto Release handles the full release cycle in one flow:
mqlaunch release
# β select option 11: Auto ReleaseSteps:
- Auto-generate changelog section from commits since last tag
- Dry run β validates VERSION, README badge, and CHANGELOG
- Live release β bumps version, tags, updates files
- Git push + push tags
- Create GitHub release
Run the release check before every push:
mqlaunch release-checkThe release check includes a repo-signal publish-readiness gate. Override the threshold with:
MQ_REPO_SIGNAL_FAIL_UNDER=14 mqlaunch release-checkSet a name that appears in every menu header:
mqlaunch nickname-set "your name"Stored in ~/.mqlaunch_nickname.
mq scanRanks CPU and memory offenders, highlights repeat offenders, and gives practical recommendations for what to close or restart.
mqlaunch doctor- checks required tools (git, brew, node, python, jq)
- validates repo state (branch, dirty tree, required files)
- evaluates workflow readiness (Git, Release, Dev, System)
- highlights issues and gives actionable recommendations
mqlaunch selftestRuns launcher smoke checks, v1 compatibility checks, and shell lint for supported bash/sh scripts.
mqlaunch demomacos-scripts/
βββ bin/ # CLI entrypoints
βββ terminal/ # menus, launchers, themes
βββ tools/ # scripts and utilities
βββ system/ # macOS helpers
βββ automation/ # workflows
βββ ai-prompts/ # AI assistant prompts
βββ ui/ # terminal UI
βββ docs/ # screenshots and documentation
βββ backups/ # workspace snapshots
- keep it simple
- structure > more tools
- optimize for real usage
- make workflows repeatable
- reduce cognitive load
See how macos-scripts is designed as a structured CLI system:
π mcamner.github.io/macos-scripts/case.html
Instead of remembering 5β10 system commands during troubleshooting:
mqlaunch system checkβ full system overview in one step
- workflow validation / health checks
- plugin-style extensions
- remote execution support
- improved onboarding
PRs welcome. If you have ideas for workflows or improvements β open an issue.
Core scripts run locally on your machine.
AI-assisted commands (ask, fix, atlas, srm) may send prompts, selected
repository context, or error text to the configured AI provider. Do not use
AI commands with secrets, private credentials, or sensitive files unless you
understand what is being sent.
Never commit API keys, tokens, or credentials. Use environment variables or ignored local files for sensitive values.
MIT


