I build local-first tools for heavy AI coding workflows and native macOS utilities — mostly things I needed myself first as a daily token burner.
- corral — run Claude Code, Codex, OpenCode, Cursor, and Pi side by side in a terminal session manager, hand off context for reviews, and continue from native iPhone & Mac clients · Python / Swift / Go
- subswap — check live quota and switch Claude Code, Codex, Kimi Code, Cursor, OpenCode, and Command Code accounts without signing in again · Rust
- agentsync — one local source of truth for
AGENTS.md,CLAUDE.md, skills, and MCP configs across 20+ tools · Go
- agent-doc-skills — navigable project knowledge base skills (
doc-init,doc-update,doc-compact) so coding agents don't rediscover your repo every session - image-skills & agent-friendly-cli-skill — specialized Agent Skills for asset processing and non-interactive CLI design
- sesskit — Python library and JSON CLI to parse and export local Claude Code, Codex CLI, and Cursor session transcripts · Python
- agent-session-control — Python library + JSON CLI + ACP adapter to launch and track local coding-agent sessions, delegating history observation to SessKit · Python
- MacResourceMonitor — menu-bar CPU monitor: see what is eating CPU, then end it
- HandySwitch — tiny menu-bar toggles (dark mode, prevent sleep, scroll direction)
- Mirror · OpenInput · NeatEditor & NeatPaste
If something here saves you time or tokens, star the repos so you can find them again later.


