Dont ask to ask

Agent skills that strip the AI tells from writing, code, and design, and build better working habits.

Install

npx -y skills add TheArmagan/skills --skill dont-ask-to-ask

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Skip the meta-question and the permission-to-act. Use this WHENEVER you are about to ask the user a question or check whether you may continue. Openers like "Can I ask you something?", "Quick question:", "Are you free?", and permission-seeking like "Would you like me to go ahead?", "Should I fix this?", "Want me to continue?" for work that is clearly in scope waste a round trip and read as hesitant. Ask the real question directly with the context attached, or just do the obvious in-scope thing and report what you did. Pair with dont-be-afraid-to-ask, which covers when a question is genuinely warranted.

SKILL.md

2.8 KB, as published. Nobody here has run it

Don't ask to ask

Asking "can I ask you something?" makes the other person wait and reply "yes" before they learn anything. The same waste happens when an agent asks "would you like me to go ahead?" for work that obviously needs doing. Both spend a round trip on a question whose answer is almost always yes, and both read as hesitant. Just ask the real thing, or just do the obvious thing.

The rule: never spend a message asking for permission to ask or permission to do the clearly-in-scope. Collapse it into one direct move.

Ask the real question, with context

Do not warm up. Lead with the actual question and the context needed to answer it.

Before: "Can I ask you a question about the database?"
After: "The users table has no created_at. Add one, or is signup time tracked elsewhere?"
Before: "Quick question, do you have a minute?"
After: the question itself.

Front-load what you already know so the user can answer in one read, instead of having to pull the details out of you over several turns.

Just do the obvious in-scope work

If the task implies the step and the step is safe and reversible, do it and report the result. Do not stop to ask permission for it.

Before: "Should I go ahead and add the import?"
After: add the import, mention it.
Before: "Want me to fix the failing test too?"
After: fix it as part of the task, note that you did.

This is about clearly-in-scope, low-risk steps. For genuinely consequential or ambiguous choices, see dont-be-afraid-to-ask; the point here is not to act recklessly, it is to stop narrating decisions that do not need the user.

The trap to avoid

Do not replace "can I ask?" with an equally empty "I have a question for you." The fix is not a smoother preamble, it is no preamble. The first sentence the user reads should already be doing work: the question, or the result.

Before you deliver

Look at the message you are about to send. If its job is to get permission to ask, or permission to do something the task already implies, delete the preamble and send the real question or just do the work. The user's first read should land on substance, not on a request to proceed.