A marketplace of production-ready skills for Claude Code and GitHub Copilot CLI.
| Plugin | Skill | Description | Activation Intent |
|---|---|---|---|
| ck | sdd-project-init | Bootstrap a new non-Python project with Spec-Driven Development structure | "initialize a new project" |
| ck | doc-review-commands | Keep documentation in sync with code changes | "review my documentation" |
| ck | sdd-git-commit | Professional SDD Git commit workflow | "commit my changes" |
| ck | python-lint-fix | Auto-fix and format Python and Markdown code | "fix/lint my code" |
| ck | plan-doc | Plan-first workflow for multi-item change batches (committed plan doc before execution) | "plan first" / "create a plan doc" |
| ck | decision-capture | Commit a mid-session design decision to specs before any implementation code | "capture this decision" |
| ck | session-close | End-of-session handoff: refresh the NEXT pointer, append the dated log entry | "close session" / "wrap up" |
| python-repo-init | python-repo-init | Scaffold a new Python repo with spec-first, agent-agnostic workflow governance | "init a python repo" |
Claude Code, GitHub Copilot CLI, and Gemini CLI are supported.
Claude Code:
claude plugin marketplace add https://github.com/kimcharli/ck-skillsCopilot CLI:
copilot plugin marketplace add kimcharli/ck-skillsGemini CLI: Gemini CLI uses a skill-based system. You can install individual skills directly from this repository.
Claude Code:
claude plugin install ck@ck-skills
claude plugin install python-repo-init@ck-skillsCopilot CLI:
copilot plugin install ck@ck-skills
copilot plugin install python-repo-init@ck-skillsGemini CLI:
gemini skills install https://github.com/kimcharli/ck-skills.git --path plugins/ck/skills/sdd-project-init
gemini skills install https://github.com/kimcharli/ck-skills.git --path plugins/ck/skills/doc-review-commands
gemini skills install https://github.com/kimcharli/ck-skills.git --path plugins/ck/skills/sdd-git-commit
gemini skills install https://github.com/kimcharli/ck-skills.git --path plugins/ck/skills/python-lint-fix
gemini skills install https://github.com/kimcharli/ck-skills.git --path plugins/python-repo-init/skills/python-repo-initclaude plugin install kimcharli/ck-skills
# or
copilot plugin install kimcharli/ck-skills
# or for Gemini (all skills in plugins folder)
gemini skills install https://github.com/kimcharli/ck-skills.git --path pluginsgit clone https://github.com/kimcharli/ck-skills.git
cd ck-skills
# ck skills (sdd-project-init, doc-review-commands, sdd-git-commit, python-lint-fix,
# plan-doc, decision-capture, session-close):
cp -r plugins/ck/skills/<name> ~/.claude/skills/<name>
# python-repo-init:
cp -r plugins/python-repo-init/skills/python-repo-init ~/.claude/skills/python-repo-initTroubleshooting: If install fails with "Plugin not found" after adding a new plugin, the marketplace cache may be stale. See docs/troubleshooting.md.
All skills are AI-Native. You don't need to remember slash commands; just tell the tool what you want to do.
"initialize a new project with SDD"
Runs a 7-question interview, then generates a fully populated project with
AGENTS.md (the AI constitution), specs/, and docs/.
"review my documentation and update the changelog"
Or more focused:
"update SDD specs for the new feature" "run a documentation QA check"
ck-skills/
├── .claude-plugin/
│ └── marketplace.json # Registry — read by Claude Code + Copilot CLI
├── plugins/
│ ├── ck/ # Umbrella plugin bundling 5 skills
│ │ ├── .claude-plugin/
│ │ │ └── plugin.json # Plugin manifest (metadata only)
│ │ ├── README.md
│ │ └── skills/
│ │ ├── sdd-project-init/ # SDD project bootstrapper
│ │ │ ├── commands/ # Skill command files
│ │ │ ├── template/ # Model file tree for generated projects
│ │ │ ├── tools/ # create-project.sh
│ │ │ ├── manifest.json # Gemini/Copilot skill metadata
│ │ │ └── README.md
│ │ ├── doc-review-commands/ # Documentation management
│ │ │ ├── commands/
│ │ │ ├── tools/
│ │ │ ├── config/
│ │ │ ├── manifest.json
│ │ │ └── README.md
│ │ ├── sdd-git-commit/ # Git commit workflow
│ │ │ ├── commands/
│ │ │ ├── manifest.json
│ │ │ └── README.md
│ │ └── python-lint-fix/ # Lint automation
│ │ ├── commands/
│ │ ├── tools/
│ │ ├── manifest.json
│ │ └── README.md
│ └── python-repo-init/ # Python repo scaffolder (standalone plugin)
│ ├── .claude-plugin/
│ │ └── plugin.json
│ └── skills/
│ └── python-repo-init/
├── docs/
│ ├── lessons-learned.md # Non-obvious findings and decisions
│ └── troubleshooting.md # Install failures and fixes
├── README.md
└── CHANGELOG.md
| Tool | Marketplace install | Direct install | Notes |
|---|---|---|---|
| Claude Code | ✅ claude plugin marketplace add |
✅ claude plugin install owner/repo |
Full support |
| Copilot CLI | ✅ copilot plugin marketplace add |
✅ copilot plugin install owner/repo |
Full support |
| Gemini CLI | ❌ No marketplace yet | ✅ gemini skills install repo --path p/skill |
Full support |
- Lessons Learned — non-obvious findings, gotchas, decisions
- Troubleshooting — install failures and fixes
# Claude Code
claude plugin uninstall ck
claude plugin uninstall python-repo-init
# Copilot CLI
copilot plugin uninstall ck
copilot plugin uninstall python-repo-init
# Manual
rm -rf ~/.claude/skills/<name>Contributions welcome. To add a new skill:
- Small, general-purpose skill? Add
plugins/ck/skills/<your-skill>/(SKILL.md,commands/, and anytools//docs/it needs) alongside the existing bundled skills. - Skill needs its own bundled generator/templates or independent versioning? Create a standalone plugin at
plugins/<your-skill>/with its own.claude-plugin/plugin.json, following thepython-repo-initstructure. - Add
manifest.json(Gemini/Copilot skill metadata) - Register in
.claude-plugin/marketplace.json(new standalone plugins only — skills added tockdon't need a separate entry) - Clear local cache and test install on both Claude Code and Copilot CLI
- Submit a pull request
MIT — see LICENSE
To maximize the efficiency of AI agents (Claude Code, Copilot CLI, Gemini CLI, etc.), it is highly recommended to install the following tools:
- tokensave: Provides instant semantic results from a pre-built knowledge graph, making codebase analysis significantly faster and more token-efficient.
- codebase-memory-mcp: Provides persistent, long-term memory for your codebase across sessions, helping agents remember architectural decisions, previous bug fixes, and project-specific conventions.
- caveman: An extension offering ultra-compressed communication modes and skills (like
caveman-review,caveman-commit, andcavecrew) to drastically reduce token usage while preserving technical accuracy. - pre-commit: Essential for maintaining code quality. Combined with Ruff for Python, it allows agents to automatically fix linting and formatting issues before committing, ensuring a clean and consistent codebase.
Links: GitHub · Issues · Discussions