Solid Go codebase, especially the lock-free CAS rate limiting and explicit crypto.ZeroKey memory wiping. Curious how you handle token billing if an upstream SSE stream drops before the final usage chunk arrives?
I currently use the built-in caffeinate command in the terminal for this exact use case (keeping AI agents or large downloads running while away).
However, caffeinate fails when I physically close the MacBook lid (unless connected to an external display/power). Looking at your architecture, it seems Lockpaw uses IOPMAssertion from IOKit, which is the same API caffeinate uses.
Does Lockpaw have any workarounds for the lid-closed sleep behavior, or is the primary advantage here combining standard sleep assertions with visual shielding/input blocking?
(Either way, solving the 'screen is on and fully accessible' problem of caffeinate is a huge win! Great app.)
Text files + Git > Vector DBs. Nice work.
I'm curious about how this scales. As the notes/ directory grows over weeks or months, reading past daily logs will eat up the context window.
Thanks for checking the list — I’m maintaining this repo and will update it based on feedback.
Quick notes:
1. This is a curated index, not runnable code — entries have one-line descriptions and tags to help scanning.
2. If something’s missing or outdated, please open an issue or submit a PR with a 1–2 line description + link — I’ll review and credit contributors quickly.
3. I’m happy to add domain-specific starter paths if people suggest them here.