By pacaplan
Trigger AI skills with 'wtf' phrases to interrupt stuck tasks, summarize progress, challenge flawed reasoning, root-cause errors, brutally review and refactor code, triage git issues, and apply quick fixes during chaotic development sessions.
npx claudepluginhub pacaplan/wtf --plugin wtfInterrupt mid-task and demand an explanation of what the agent is doing and why. Use when the user asks "wtf are you doing", "what's the plan", "explain yourself", "where are you going with this", "stop and explain", or "what are you working on".
Challenge something the agent just said and force a genuine re-examination. Use when the user asks "wtf are you thinking", "that's wrong", "push back", "are you sure", "that doesn't sound right", "re-examine", or "double check that".
TL;DR summary of recent verbose agent activity. Use when the user asks "wtf did you say", "what just happened", "summarize what you did", "catch me up", "tldr", or "what did you do".
Skip the diagnosis monologue and go straight to fixing the current problem. Use when the user says "wtf fix it", "just fix it", "make it work", "fix this", or "stop talking and fix it".
Brutally honest code review followed by a refactor. Use when the user asks "wtf is this", "this code is terrible", "refactor this", "clean this up", "code review", "this is a mess", "look at this code", or "review this".
Triage everything that's broken and deliver a prioritized action plan. Use when the user asks "wtf should I do", "where do I start", "I'm stuck", "what's next", "help me prioritize", "I don't know what to do", or "triage this".
Brutally honest review of your own recent diffs and commits before opening a PR. Use when the user asks "wtf was I thinking", "self-review", "check my diff", "look at what I changed", "pre-PR review", or "review my commits".
Root cause analysis that traces the chain of causation, not just the symptom. Use when the user asks "wtf went wrong", "why is this broken", "debug this", "what happened", "why did this fail", "trace this error", or encounters an error, stack trace, test failure, or build failure.
Evaluate a crazy idea and make an honest case for why it might actually work. Use when the user asks "wtf why not", "should I build this", "crazy idea", "hear me out", "this is stupid but", "would this work", "devil's advocate", or pitches something unconventional and wants a real assessment.
Quick acknowledgement and redirect when the user says "wtf" without context. Use when the user says "wtf" as a standalone reaction or expression of frustration without a specific task.
A Claude Code plugin for when things go sideways.
Ten debugging, explanation, and code review skills for the moments when something just broke, you can't remember what you were doing, or you need a second opinion before shipping.
It started with this:

So I said (jokingly, or so I thought):

And then I thought: WTF - why not?
Are these skills well thought out? Not really. But are they useful? Maybe.
| Command | What it does |
|---|---|
/wtf:are-you-doing | Interrupt mid-task and demand an explanation of the plan. |
/wtf:are-you-thinking | Push back on something Claude just said. Forces a genuine re-examination. |
/wtf:did-you-say | TL;DR of a long autonomous agent chain. The "I stepped away for coffee" button. |
/wtf:fix-it | Skip the lecture. Just make it work. |
/wtf:is-this | Brutally honest code review, followed by a refactor. |
/wtf:should-i-do | Triage everything that's broken and give a prioritized action plan. |
/wtf:was-i-thinking | Self-review your own recent diffs and commits before opening a PR. |
/wtf:went-wrong | Root cause debugging. Traces the chain of causation, not just the symptom. |
/wtf:why-not | Evaluate an unconventional idea and make an honest case for why it might work. |
/wtf:wtf | Quick acknowledgement and redirect when you just say "wtf" without context. |
In Claude Code, add the wtf marketplace and install the plugin:
claude plugin marketplace add pacaplan/wtf
claude plugin install wtf
claude plugin marketplace update wtf
claude plugin update wtf@wtf
All skills accept optional arguments for context:
/wtf:went-wrong it started failing after the last commit
/wtf:is-this this class is way too long
/wtf:was-i-thinking
Or just type "wtf" when something breaks. The plugin will know what to do.
MIT
Capture mistakes and improvement opportunities with automatic 5 whys analysis
Share bugs, ideas, or general feedback.
Analytical thinking tools - dialectic analysis, bias elimination, codex consultation
Developer experience essentials: GitHub Actions debugging, conversation cloning/half-cloning, context handoffs, and Reddit research via Gemini CLI
Codex CLIにセカンドオピニオンを求めるスキル。「codexと相談して」「codexに聞いて」で発動する。
Read The Fucking Prompt — finds the strongest user reaction to an AI instruction-following failure in a chosen session, reconstructs the triggering assistant output, and renders a shareable terminal-style PNG.
Automatic code review, adversarial review, and rescue via Codex.