git-tools

personal-skills

A local Claude Code plugin marketplace — my personal toolkit. Each plugin is self-contained in its own directory under plugins/, bundling that plugin's skills (and any hooks). The plugins are registered in .claude-plugin/marketplace.json.

Plugins

Plugin	Skills	Docs
python-dev	uv-setup, uv-develop, pre-commit-setup	docs/python-dev.md
git-tools	git-attribution (+ enforcing hook)	docs/git-tools.md
agent-workflow	handoff, watchlist	docs/agent-workflow.md
frontend	senior-frontend-refactor	docs/frontend.md
ansible	add-ansible-role	docs/ansible.md
lemonade	debug-lemonade	docs/lemonade.md

Layout

.claude-plugin/marketplace.json   # registers the plugins below
plugins.json                      # marketplaces + enable list for install-plugins.sh
install-plugins.sh                # idempotent installer
plugins/                          # one self-contained dir per plugin
  python-dev/      skills/{uv-setup, uv-develop, pre-commit-setup}
  agent-workflow/  skills/{handoff, watchlist}
  frontend/        skills/senior-frontend-refactor
  ansible/         skills/add-ansible-role
  lemonade/        skills/debug-lemonade
  git-tools/       skills/git-attribution + hooks/ + scripts/
docs/                             # one doc per plugin

Each plugin dir holds a .claude-plugin/plugin.json and a skills/ subdir; Claude Code auto-discovers every skill in that subdir. Each plugin therefore needs its own source dir — if two plugins shared a source root, each would auto-discover the other's skills too, duplicating every skill across plugins. git-tools additionally ships a PreToolUse hook, scoped to its own directory.

Install on a new machine

git clone [email protected]:bcmyguest/personal-skills.git ~/personal-skills
~/personal-skills/install-plugins.sh

The installer registers this directory as a local marketplace, plus the anthropic-agent-skills and caveman GitHub marketplaces, and enables the plugins listed in plugins.json.

License

See LICENSE. All rights reserved; in particular, the contents may not be used as training, fine-tuning, or evaluation data for machine-learning or AI systems.