A curated collection of Claude Code Skills used by ShruggieTech across client engagements, internal tooling, and consulting work. Skills here encode consistent output formats, workflow automation, and house-style conventions so the same standards apply whether Claude Code is running on a Proxmox host, a personal workstation, or a collaborator's laptop.
Skills are modular folders containing a SKILL.md instruction file plus optional supporting assets. They load the same way whether you use Claude at claude.ai in a browser, the Claude desktop app, or the Claude Code CLI: a skill triggers either automatically (when your prompt matches its description) or explicitly via /skill-name. See the official skills documentation for the full model.
Actively maintained. Treat the main branch as the authoritative source; tagged releases are cut on a rolling basis when meaningful changes ship.
Skills in this repo are distributed two ways. Pick whichever matches how you use Claude.
If you use Claude at claude.ai in a browser, or the Claude desktop app on macOS or Windows, this is the path you want.
- Open the latest release and download the zip for each skill you want. Files are named
<skill-name>-vX.Y.Z.zip(for example,shruggie-html-v1.0.0.zip). - In Claude, open Settings and find the skill upload UI. Recent builds put it under Capabilities → Skills, though the exact label moves as Claude evolves; the in-app search at the top of Settings finds it quickly. Upload the zip as-is. Do not unzip it first.
- The skill is now available in your conversations. It triggers automatically on prompts that match its description, or invoke it explicitly with
/skill-name.
One zip per skill, one upload per skill. Repeat the upload for each skill you want available.
If you want to compare hashes before uploading, every release ships a SHA256SUMS.txt next to the zips on the Releases page.
If you run the Claude Code CLI on your own machine, you have a second option: symlink this repo into Claude Code's skill directory so git pull keeps you on the latest version without re-uploading anything. (You can still use the zip approach above if you prefer; the CLI honors uploaded skills the same way the browser does.)
The symlink path below is for the Claude Code CLI only. Symlinking into
~/.claude/skills/has no effect on the Claude browser or desktop app; those load skills from your Claude account's storage, not from your filesystem.
Personal skills directory by platform:
| Platform | Personal Skills Path | Project Skills Path |
|---|---|---|
| Linux | ~/.claude/skills/<name>/ |
<project>/.claude/skills/<name>/ |
| macOS | ~/.claude/skills/<name>/ |
<project>/.claude/skills/<name>/ |
| Windows 11 | %USERPROFILE%\.claude\skills\<name>\ |
<project>\.claude\skills\<name>\ |
Personal skills are available across all your CLI projects. Project skills are scoped to one repository and travel with it through version control.
git clone https://github.com/shruggietech/skills.git ~/.shruggietech-skills
cd ~/.shruggietech-skills
./scripts/install.shSymlinks on Windows 11 require either an elevated PowerShell session or Developer Mode enabled.
git clone https://github.com/shruggietech/skills.git "$env:USERPROFILE\.shruggietech-skills"
Set-Location "$env:USERPROFILE\.shruggietech-skills"
.\scripts\install.ps1Both installers create symlinks rather than copies, so a git pull updates every linked skill in place.
If a skill should only apply to a single repository, link it directly into that project's .claude/skills/ directory.
Linux and macOS:
mkdir -p .claude/skills
ln -s ~/.shruggietech-skills/skills/<skill-name> .claude/skills/<skill-name>Windows 11 (PowerShell):
New-Item -ItemType Directory -Force -Path .claude\skills | Out-Null
New-Item -ItemType SymbolicLink `
-Path ".claude\skills\<skill-name>" `
-Target "$env:USERPROFILE\.shruggietech-skills\skills\<skill-name>"Commit the symlink (or a copied folder) so collaborators get the same behavior.
Claude Code watches skill directories. Adding, editing, or removing a skill takes effect in the current session without restarting, as long as the parent skills directory existed when the session started.
skills/
├── README.md # This file
├── LICENSE # Apache 2.0
├── NOTICE # Required attribution notices
├── CHANGELOG.md
├── CONVENTIONS.md # House-style rules every skill must follow
├── CONTRIBUTING.md # How to propose or add a skill
├── TODO.md # Running list of engineering follow-ups
├── scripts/
│ ├── install.sh # CLI symlink helper for Linux and macOS
│ ├── install.ps1 # CLI symlink helper for Windows 11
│ ├── release.sh # Cut a tagged release (Linux and macOS)
│ └── release.ps1 # Cut a tagged release (Windows 11)
└── skills/
├── _template/ # Starting point for new skills
│ └── SKILL.md
└── <skill-name>/
├── SKILL.md # Required: frontmatter + instructions
├── README.md # Optional: human-facing notes
├── assets/ # Optional: templates, references, design tokens
└── scripts/ # Optional: executable helpers (bash, python, etc.)
Each skill is a self-contained folder. SKILL.md is the only required file; everything else is optional and referenced from the body of SKILL.md when needed. Supporting files load lazily, so a skill can carry substantial reference material without inflating idle context.
Every skill in this repo follows a shared set of formatting and style rules so output stays consistent regardless of which skill is active. The full ruleset is in CONVENTIONS.md. Summary:
- UTF-8 encoding without BOM, no trailing whitespace
- No em-dashes in any prose output (use parentheses, commas, or standard hyphens)
- Markdown follows audience-conditional formatting (AI-only vs human-facing)
- Plans and logs are sequenced chronologically
- Every
SKILL.mdfrontmatter declares at minimum:description(with clear trigger phrases) and an explicitdisable-model-invocationvalue
New skills are welcome from ShruggieTech operators and trusted collaborators. See CONTRIBUTING.md for the full workflow. Quick version:
- Branch off
main - Copy
skills/_template/toskills/<your-skill-name>/ - Fill in the frontmatter and body
- Verify the skill triggers correctly with a local
claudesession - Update
CHANGELOG.md - Open a pull request
Skill names use lowercase letters, numbers, and hyphens only (max 64 characters per the Agent Skills spec).
Apache License, Version 2.0. See LICENSE and NOTICE.
Copyright 2026 Shruggie LLC (DBA ShruggieTech)
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Individual skills may bundle scripts or assets under other compatible licenses; check the relevant skill's README.md if present.
- Claude Code: Extend Claude with skills - canonical reference for skill structure, frontmatter, lifecycle, and sharing
- Claude Code overview - top-level entry point for all Claude Code documentation
- Claude Code commands reference - built-in commands and bundled skills (
/simplify,/debug, etc.) - Claude Code plugins - bundle skills and other extensions into installable packages
- Claude Code subagents - delegation patterns, including preloading skills into subagents
- Claude Code hooks - automate workflows around tool and skill lifecycle events
- Claude Code permissions - control tool and skill access at the workspace and managed-settings level
- Claude Code settings reference - including
skillOverrides,skillListingBudgetFraction, and managed settings - Agent Skills open standard - cross-tool spec that Claude Code skills implement
- awesome-claude-code - curated index of skills, hooks, plugins, and orchestrators
- Thariq Shihipar: Using Claude Code, The Unreasonable Effectiveness of HTML - rationale and examples for HTML-as-output workflows
- Anthropic Engineering Blog - deeper dives on Claude Code internals and patterns