Simple Skills

╔══════════════════════════════════════════════════════════╗
║                                                          ║
║   ____  _                 _         ____  _    _ _       ║
║  / ___|(_)_ __ ___  _ __ | | ___   / ___|| | _(_) |___   ║
║  \___ \| | '_ ` _ \| '_ \| |/ _ \  \___ \| |/ / | / __|  ║
║   ___) | | | | | | | |_) | |  __/   ___) |   <| | \__ \  ║
║  |____/|_|_| |_| |_| .__/|_|\___|  |____/|_|\_\_|_|___/  ║
║                     |_|                                  ║
║                                                          ║
║     think → design → plan → execute → review → done      ║
║                                                          ║
╚══════════════════════════════════════════════════════════╝

Agent kit for work that ships
think → design → plan → execute → review → done

What · Install · Quick start · Profiles · Model · Thinking · Start here

What is this?

Simple Skills is an agent kit that gives AI coding assistants structured workflows for real work. Instead of vague prompts, you get:

• 66 skills across 5 profiles (core, office, BA, frontend, backend)
• 12 thinking methods for better decision-making
• Session management with automatic git integration
• Artifact validation to ensure quality
• Smart install that handles conflicts gracefully

Perfect for teams who want AI assistance that actually ships.

Install

The easiest way to install and manage Simple Skills across all operating systems (macOS, Linux, Windows) is using pipx or uv.

1. Install the global CLI tool

pipx install simple-skills
# or: uv tool install simple-skills

2. Install skills into your project

Navigate to your project directory and run:

cd your-project
sk install

(Note: Use sk install --agent claude to install into .claude instead of the default .agents directory).

Available Commands

# Install everything (replaces existing directory)
sk install

# Update your skills from upstream (keeps your custom skills safe)
sk update

# Check if your project has all required configuration and docs
sk doctor

curl -fsSL https://raw.githubusercontent.com/truongnat/simple-skills/main/install.sh -o install.sh
bash install.sh install

iwr -useb https://raw.githubusercontent.com/truongnat/simple-skills/main/install.ps1 -OutFile install.ps1
.\install.ps1 install

Quick start

After installation, run the init skill once to set up your project:

# In your AI assistant, say:
"Use the init skill to set up this project"

Then choose your path based on the task:

Lite/Full paths use Step ledger (track each workflow step) and Spec quality (validate artifacts before moving on) to ensure nothing falls through the cracks.

Check progress anytime:

bash .agents/tools/session/session.sh status

CLI reference

The sk CLI is beautifully simple:

# Install everything into the default .agents folder
sk install

# Install for a specific agent provider (e.g. into .claude)
sk install --agent claude

# Update your skills from upstream (keeps your custom skills safe)
sk update --agent claude

# Check if your project has all required configuration and docs
sk doctor --agent claude

That's it. No complicated flags needed.

Mental model

flowchart LR
  K["Kit · .agents/"] --> W["Work · .agent-work/"]
  K --- S[skills · tools · policy · thinking]
  W --- T[sessions · memory · nested git]

Key principle: Progress lives in TASKS.md + session.sh status. Artifacts only under sessions. Blocking changes require confirmation.

Thinking methods

These are session-wide principles, not report section titles. They guide every decision:

Outcome-first → IPO → Make-explicit → SSOT → Small-batch → Feedback loop
→ Default path first → Reversible → Standardize → Handoff → Evidence
→ Bottleneck → (5W1H) → Vital few

How to use: Fold these into Goal / Verify / Handoff sections. No separate OUTCOME.md or HANDOFF.md files needed.

Full details: docs/thinking/README.md

HTML · Docs · Develop

• Design system: DESIGN_SYSTEM.md · seed VISUAL_DECISION.template.html
• Docs map: docs/README.md
• Getting started: START_HERE.md
• What's next: WHAT_NEXT.md

Development setup

pip install -e ".[dev]"
python scripts/validate_skills.py && pytest -q

Python ≥ 3.11 · MIT License
github.com/truongnat/simple-skills