Thinking Agent Skills

This repository contains Agent Skills for structured reasoning, mental models, decision frameworks, systems thinking, communication, and problem analysis.

Skills are markdown files that give AI agents operational guidance for applying a specific thinking tool at the right moment.

Repo Layout

.
├── README.md
├── LICENSE
├── docs/
│   └── authoring.md
├── templates/
│   └── skill-template/
│       └── SKILL.md
└── skills/

Add a New Skill

1. Create a new directory inside skills/ using a lowercase kebab-case name.
2. Copy templates/skill-template/SKILL.md into that directory.
3. Update the name field so it exactly matches the directory name.
4. Replace the placeholder description with a specific trigger-oriented description.
5. Fill in the instructions with the workflow you want agents to follow.
6. Add optional references/, scripts/, or assets/ folders only when needed.

Example:

skills/
└── your-skill-name/
    ├── SKILL.md
    ├── references/
    ├── scripts/
    └── assets/

Authoring Rules

• Keep skill names lowercase and hyphenated.
• Keep SKILL.md focused and practical.
• Put long reference material in references/ instead of bloating SKILL.md.
• Write descriptions so an agent can tell when to activate the skill.
• Prefer reusable workflows over product-specific prompts.

See docs/authoring.md for a fuller checklist.

Available Skills

Skill	Description
abstraction-laddering	Reframe problems by moving between higher-level why questions and lower-level how questions.
balancing-feedback-loop	Identify stabilizing mechanisms that push systems back toward a goal or desired level.
concept-map	Visualize relationships between entities in a concept, domain, or system.
confidence-determines-speed-vs-quality	Decide how much polish, validation, or speed a product decision needs.
conflict-resolution-diagram	Resolve opposing demands by uncovering underlying needs and a shared goal.
connection-circles	Map cause-and-effect relationships and identify feedback loops in a system.
cynefin-framework	Classify situations by domain complexity and choose an appropriate response.
decision-matrix	Choose between options by scoring them against weighted criteria.
eisenhower-matrix	Prioritize work by urgency and importance.
first-principles	Break complex problems into fundamental truths and rebuild from the basics.
hard-choice-model	Classify a decision by impact and ease of comparison.
iceberg-model	Move from visible events to patterns, structures, and mental models.
impact-effort-matrix	Prioritize tasks by expected impact and required effort.
inversion	Reason backward from failure, bad outcomes, or the opposite of success.
ishikawa-diagram	Identify possible root causes by organizing contributing factors.
issue-trees	Break complex problems into mutually exclusive and collectively exhaustive branches.
ladder-of-inference	Inspect selected data, interpretations, assumptions, beliefs, and actions.
minto-pyramid	Structure communication from conclusion to supporting points and details.
ooda-loop	Make faster decisions in changing environments by observing, orienting, deciding, and acting.
productive-thinking-model	Solve defined problems creatively through a structured six-step process.
reinforcing-feedback-loop	Identify dynamics that amplify growth, decline, or compounding effects.
second-order-thinking	Examine consequences beyond immediate effects.
situation-behavior-impact	Give clear feedback by naming the situation, observed behavior, and impact.
six-thinking-hats	Examine a decision or problem from multiple perspectives.

Validation

The Agent Skills spec recommends validating each skill with skills-ref:

skills-ref validate ./skills/your-skill-name

Spec reference: https://agentskills.io/specification