Claude Code thinking skills
A catalog of 28 mental-model skills for agents: diagnosis, decisions, systems, risk, strategy and idea generation
- 1,729 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skills consist only of instructions and reasoning procedures
Install
- Works in
Manual install
/plugin marketplace add tjboudreaux/cc-thinking-skills
/plugin install thinking-skills@thinking-skills-marketplaceAccess and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skills consist only of instructions and reasoning procedures
This is third-party code. Review the repository files before installing.
What it does
The set gives the agent 28 portable skills, each defining a clear reasoning procedure for one mental model. Diagnosis skills include the scientific method, five whys, the Kepner-Tregoe method and systems thinking. Decision skills include reversibility, opportunity cost, probabilistic thinking, second-order effects and steel-manning. Creation and improvement skills include first principles, TRIZ, via negativa, thought experiments and jobs-to-be-done. Risk and execution skills include the pre-mortem, red team, margin of safety, theory of constraints and the OODA loop. A separate router skill helps pick the right model or honestly return that none is needed: it outputs nothing, one primary skill, or up to three complementary ones. Each skill states when to use it and when not to. It installs as a Claude Code plugin, via npx skills or by copying folders, and also works in Copilot, Codex, Cursor and other Agent Skills-compatible tools.
Who it is for. For product managers, founders and developers who want a ready toolkit of structured reasoning to work through problems with an agent.
Good fit when
- A problem needs a clear reasoning frame rather than a generic answer
- You are unsure which mental model to use: the router suggests one
- You want to review a decision or launch plan, or diagnose a failure by a procedure
Not a fit when
- The task is routine with no analytical unknown
- The right model is obvious: call that skill directly without the router
Example request
Use thinking-pre-mortem to find where our launch plan could failLimitations
The skills give a reasoning procedure, not a finished answer: output quality depends on the inputs and the agent's judgment. By the project's own audit all skills are manual-only, there is no automatic trigger mode, and the authors claim no measured accuracy gain and ask not to present provisional numbers as a proven result. Some checks and evals in the repository need Node.js and a separate CLI. MIT licensed.
How to disable. Remove the plugin via /plugin, or delete the skill folders from the agent's skills directory, for example .claude/skills/ or ~/.claude/skills/.
FAQ
Where do I start if I do not know which model to use?
With the thinking-model-router skill. It assesses domain and problem type and returns nothing, one primary skill, or up to three complementary ones.
Do the skills trigger on their own?
No. By the project audit all skills are invoked manually by name; there is no automatic mode.
Discussion
No questions or reviews yet.
Related
SkillsA library of 229 skills for work and life: sales, marketing, finance, engineering, design and parallel agent runs
A collection of skills for decisions and projects: strategy, market research, thinking, data analysis, docs and slides.
A set of skills that turn an AI agent into a tutor built on cognitive science, from school to career and social skills
Ten skills for Claude Code and Codex: a router diagnoses a stuck idea and runs the matching lateral thinking technique
Second Brain Skills
Claude Code Second Brain Skills
A set of Claude Code skills for knowledge work: brand and voice, an MCP client, PPTX slides, Remotion video, SOPs and a skill creator
A set of 13 SKILL.md skills: research, a product doc, decks, articles and audits for any compatible agent
README in short
The README presents a catalog of 28 portable structured-thinking skills for Claude Code, Copilot, Codex, Cursor and other compatible tools. The skills are grouped by task: diagnosis, decisions, creation and improvement, risk and execution, plus a router for choosing the right model. Installation is via npx skills, as a Claude Code plugin with /plugin install thinking-skills@thinking-skills-marketplace, or by copying folders from skills/. The README asks you to read the catalog audit before any performance claims: all skills are manual-only, there is no measured accuracy gain, and a provisional figure sits below the utility threshold. MIT licensed.
README badge
Original SKILL.md
--- name: thinking-model-router description: When unsure which thinking skill fits, map domain and problem type, then return NONE or one primary skill by default (at most three complementary). disable-model-invocation: true --- # Model Router **Core rule:** Prefer NONE or one primary skill. Route by mechanism fit, not habit. Combine only when roles are distinct and necessary. ## When to Use - The right thinking skill is unclear and you would otherwise guess or stack tools. - Several catalog skills seem plausible and you need a single primary (or explicit NONE). - High-stakes work where a wrong frame is costly and a quick domain by type match helps. ## When NOT to Use - The match is already known or obvious, invoke that skill directly. - The task is routine implementation with no analytical unknown, reason directly (NONE).