How to use the Explain It Plainly agent skill (wtf)

What the wtf skill makes the agent do

The skill tells the agent to explain a target the way it would to a smart colleague who was not in the room. It opens with the main point and why it matters to you, then says what happened, what state things are in, and anything you need to decide or do.

The explanation is meant to be shorter than the source. The agent is told to keep failures, caveats, open questions, and requests, and to avoid sounding more certain than the source. If the source is vague or contradicts itself, the agent should say so.

• With nothing passed, it explains the agent's most recent message.
• With a file path or link, it reads that and explains it.
• With a pasted passage, it explains that passage.
• With a short pointer such as "the migration part" or "step 3", it explains only that part of the conversation.
• If the target is unclear, it asks one short question instead of guessing.

Install the wtf skill from the ZIP

Download the ZIP for skills/wtf from the directory page, then unzip it. The skill folder contains SKILL.md and agents/openai.yaml.

Put the unzipped skill folder in one of two places. For a personal install across projects, use ~/.claude/skills/. For a single project, use the .claude/skills/ folder inside that project.

If you use the Claude apps instead, you can upload the skill under Settings → Capabilities → Skills.

• Personal: ~/.claude/skills/
• Project only: .claude/skills/ inside your project folder
• Claude apps: Settings → Capabilities → Skills, then upload

Trigger the skill with a natural request

The skill is set to manual invocation, so the agent will not run it on its own. You start it by typing the skill's name, wtf, in your request, and you can add a file path, URL, pasted text, or a description of which part lost you.

The skill's argument hint is: [file path, URL, pasted text, or which part lost you; blank for the last message]. Anything you add after wtf is treated as the target.

• Type "wtf" alone to get the last message explained.
• Type "wtf step 3 of your plan" to focus on one part.
• Type "wtf" followed by a file path or link to have that document explained.
• Paste a passage after "wtf" to have just that text explained.

Examples of when to use it

The skill fits moments when a reply from an agent or a developer's update made sense to its author but not to you. You still need the outcome, so the goal is to get the gist and any action items in plain words.

Use it for status reports and reviews, where the key decisions matter more than the detail.

• A coding agent ends with a dense summary and you need to know whether anything is waiting on you.
• A developer sends a link to a jargon-heavy design document and you need its key decisions.
• A pasted status update from an engineering team that you have to respond to.
• You want only the migration section of a long reply explained.

What you need and what the skill does not do

The listing does not name any required tools. The skill does not run scripts, but it does use the network, which is how it reads links. It is designed for explanation only, so it explains what the source says and cannot verify facts beyond it.

For other kinds of help, the skill points elsewhere. Use ce-explain to investigate how or why code works, and ce-noslop to rewrite prose for style or tone.

• Requires: nothing listed
• Runs scripts: no
• Uses network: yes (for example, to read a URL you pass in)
• Not for code investigation (use ce-explain) or style rewrites (use ce-noslop)

Safety notes and tips

Because the skill uses the network, a link you pass in may be fetched by the agent. Only pass links you are comfortable having it read.

For the most useful result, pass the exact file name, command, or value you will need to act on, and say which part confused you. The skill keeps exact file names, commands, and values so you can use them afterwards.

• Name the specific part that lost you to get a tighter explanation.
• Check any decision the explanation flags against the original source before acting on it.
• If the explanation says the source is vague, treat that as an open question rather than a settled fact.