An agentic skills framework & software development methodology that works.
Thinking skill for Claude Code, Codex, and other agents
Overthink is a skill pack that makes an agent apply a deliberate thinking process before it responds. It is built around eight modes of thought, which the agent combines based on the prompt so it can stay thorough without turning every task into a long discussion.
Builders who use Claude Code, Codex, Cursor, Copilot, or other skill-aware agents and want more careful reasoning.
You can get fuller answers on planning, design, naming, tradeoffs, and reviews instead of the first obvious reply.
What it does
Eight thinking modes
Defines model, create, diagnose, verify, evaluate, decide, taste, and synthesize as separate ways the agent can think.
Prompt-sensitive process
Builds a thinking process from the modes that fit the current prompt, so the agent can stay focused on the task.
Automatic lightweight runs
Lets agents trigger the skill automatically on risky or ambiguous tasks while keeping normal work fast.
Manual `/overthink` use
Provides a slash command pattern you can place before a prompt when you want the full process.
How to get it
- 1Run
npx skills add s0md3v/overthink
README
overthink
Force the agents to think things through.
No more:
- short sightedness
- "you are absolutely right!"
- being generic when it should be creative
- missing the point and/or details
- failing to assess consequences, tradeoffs, hidden assumptions etc.
- relying on old info when it should search online
This skill addresses this by defining 8 "modes" of thinking and then creating a thinking process by combining the modes that matter for the prompt. The agent then follows that process to be thorough.
Install
npx skills add s0md3v/overthink
Works with Codex, Claude Code, Cursor, Copilot, OpenCode, and other agents that support skills. Restart your agent harness if it does not pick up newly installed skills automatically.
You can also directly download overthink.zip from releases and extract it in the skills directory of your choice.
Use
Put /overthink before your prompt:
/overthink suggest some concepts for the landing. it needs to stand out but also be usable and professional.
/overthink make the plugin archtecture more secure
/overthink look at the state of the project, read the docs+notes and tell me what should we do next
/overthink how do we make this interface simpler to use without hiding the functionality?
/overthink should we port it to rust?
Agents can also trigger overthink automatically when a task looks risky or ambiguous. Automatic runs are lighter so normal work does not turn into a seminar. Use /overthink when you want the full process.
Modes
The skill has eight modes: model, create, diagnose, verify, evaluate, decide, taste, and synthesize.
Use it for debugging, research, reviews, tradeoffs, planning, design, naming, writing, and any prompt where the first obvious answer is probably not good enough. Skip it for simple facts, mechanical edits, and one-step tasks.
Credits
- Hypatia's Potrait, drawn by Jules Maurice Gaspard
- "How we think" by John Dewey
- "The fixation of belief" by C. S. Peirce
- "How to solve it" by George Polya
- "Strong Inference" by John R. Platt
- "Cargo cult science" speec by Richard Feynman
- "Reflections on the right use of school studies" essay by Simone Weil
- "Critique of judgment" by Immanuel Kant
- "Notes on the synthesis of form" by Christopher Alexander
- Juxtopposed youtube channel
- victors1681/3dbrain for the brain visualization
- works with component adopted from pbakaus/impeccable
Files in the repo
- .github
- site
- skill
- .gitignore
- README.md
Discussion (0)
Ask about usage, or say what you built with itSign in to join the discussion.
No comments yet. Be the first to say what this is good for.
More skills

Turn any codebase, with its docs, SQL schemas, configs, and PDFs, into a queryable knowledge graph. A /graphify skill for Claude Code, Cursor, Codex, and Gemini CLI: local deterministic AST parsing, every edge explained, no vector store.
Topic in, narrated explainer video out. A Claude Code / Codex skill that turns any topic into a black-canvas motion-graphics explainer video with TTS voiceover, subtitles and a chapter progress bar. Chinese or English; every frame drawn in code with Remotion.
Public repository for Agent Skills
Open-source AI job search: scan job portals, evaluate listings into a structured A-H report with a global 1-5 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)

Production-grade engineering skills for AI coding agents.