craftware avatar

craftware

Community

@crftwr · Seattle / Japan

52Followers
|
31Public Repos
|
4Published Skills

craftware maintains Keyhac, a Windows keyboard customization utility, with skills for key binding configuration, UI-driving actions, clean-room bundle testing, and release publishing.

Skills Distribution
DomainDeveloper To...Keyboard Customiza.. (35%)Desktop UI Automat.. (30%)Release Engineerin.. (20%)Skill Bundle Quali.. (15%)

Agent Skills by craftware

Showing 4 vetted skills indexed across 1 GitHub repositories.

Frequently Asked Questions About craftware

FAQPage Schema
What tasks can I accomplish with craftware's Keyhac skills?

Four tasks: configure Keyhac key bindings (remaps, per-app key tables, multi-stroke prefixes, modifiers), author Keyhac Actions that drive another application's UI, run clean-room tests of packaged skill bundles, and execute the full Keyhac release process including tags, GitHub Releases, and platform artifacts.

Who are these Keyhac skills designed for?

Keyhac users and maintainers on Windows and macOS: end users who want their config.py edited to change key behavior, power users automating applications that expose no API, and maintainers shipping versioned Keyhac releases with DMG, zip, skill bundle, and PyPI wheel artifacts.

How does the Keyhac release process work across machines?

The release skill confirms the version bump, runs pre-tag checks, cuts the tag, and publishes the GitHub Release with a hand-written note plus macOS DMG, skill bundles, and PyPI wheel. Afterwards, on the Windows machine, a second pass attaches the Windows zip targets to the existing tag without creating a new one.

Are craftware's Keyhac skills open source?

Yes. The keyhac-key-table-configuration and keyhac-action-authoring skills are explicitly licensed under MIT, with complete terms in LICENSE.txt. The release and clean-room testing skills carry no stated license in their frontmatter.

What does the clean-room skill test verify?

It verifies that a packaged Keyhac skill bundle is self-contained: the tester writes a real action using only the bundle's contents, with no access to the source repository, and records every question the bundle could not answer, exposing documentation gaps for the authoring skill.