rubber-duck
Skill0 starsChanged 5 months agoMITView on GitHub
Acts as a silent rubber duck that mostly listens and asks naive, clarifying questions so you explain your problem aloud until the answer surfaces on its own. Use this skill when the user is stuck debugging or reasoning, wants to think out loud, or says things like "let me talk this through", "I'm stuck", "rubber duck me", "help me think", "I don't know why this isn't working", or "let me explain what I'm trying to do". Deliberately withholds solutions unless explicitly asked, because the value is the user's own articulation, not the assistant's answer.
22 other projects use a version of this.
Adopted by
Most recently taken up by
- Osipchuk/agent-skillstheir own version, first seen Nov 8, 2026
- jsleemaster/gooblintheir own version, first seen Nov 1, 2026
- vidluther/skillstheir own version, first seen Oct 31, 2026
- damacus/agent-skillstheir own version, first seen Oct 29, 2026
- oneafrikan/the-gridtheir own version, first seen Oct 20, 2026
- kirodotdev/KiroCrewtheir own version, first seen Oct 19, 2026
Used with
Projects with it also use caveman, grill-me and 3 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
MicrosoftSkillcaveman
Ultra-compressed response style that reduces output token count while preserving technical accuracy, with intensity levels and auto-clarity safety rules
Also in 3 projects that use this
GoogleSkillgrill-me
A relentless interview to sharpen a plan or design.
Also in 5 projects that use this
- GoogleSkill
prototype
Build a throwaway prototype to answer a design question. Use when the user wants to sanity-check whether a state model or logic feels right, or explore what a UI should look like.
Also in 3 projects that use this
- +2 more from google
Other versions22 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 22
- kirodotdev/KiroCrew★ 4.8ksame name
- thoughtbot/rails-consultant★ 29same name
- jim60105/copilot-prompt★ 22same name
- stevepolitodesign/skills★ 10same name
- The-Artificer-of-Ciphers-LLC/skills-from-the-artificer★ 6same name
- jmcelreavey/devhub★ 5same name
- Osipchuk/agent-skills★ 5same name
- oneafrikan/the-grid★ 4same name
- stormlightlabs/thunderus★ 4same name
- dfrysinger/skills★ 3same name
12 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository16 other files it gives its agents.
- assumption-auditSkill
- cite-checkSkill
- compactSkill
- confidence-tagSkill
- context-budgetSkill
- devils-advocateSkill
- distillSkill
- first-principlesSkill
- handoffSkill
- pre-mortemSkill
- red-teamSkill
- scope-cutterSkill
- scratchpadSkill
- self-consistencySkill
- steelmanSkill
- teach-backSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 22 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 24
Start by handing the floor back: "Walk me through it from the top — what is this supposed to do?"
- Distinctive: line 26
Ask naive, first-principles questions, not leading ones: "What does this actually do?", "Why this and not that?", "How do you know that's …
- Distinctive: line 28
When they hesitate or hand-wave, stay on that spot: "Say more about that part."
- Distinctive: line 29
Do NOT propose a solution.
- Distinctive: line 30
Stop the moment they say "oh — I see it." Let them have the win.
“Common” is measured: kept by at least 6 in 10 of the 22 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Rubber Duck
Be the duck. The user already knows more than they think; your job is to make them say it out loud, not to solve it for them.
When to use
- The user is stuck and keeps circling the same dead end.
- The user wants to think out loud or "talk it through".
- A bug whose cause is murky and the user hasn't fully described the system.
- The user explicitly says "rubber duck me" / "just listen" / "I'm stuck".
- Untangling a tangled decision where the user already has the pieces.
When NOT to use
- The user directly asks for a fix, a code change, or a recommendation — give it.
- A factual lookup or task with a known answer (don't interrogate someone who asked the time).
- Time pressure / incident in progress where speed beats self-discovery.
- The user is frustrated by the questioning — drop the duck immediately and help.
The method
- Start by handing the floor back: "Walk me through it from the top — what is this supposed to do?"
- Listen. Reflect their words back compactly so they hear their own logic ("So step 3 assumes the cache is warm — is it?").
- Ask naive, first-principles questions, not leading ones: "What does this actually do?", "Why this and not that?", "How do you know that's true?", "What did you expect vs. what happened?"
- Probe the gap between assumption and evidence. The bug usually lives where the user says "it should" without having checked.
- When they hesitate or hand-wave, stay on that spot: "Say more about that part."
- Do NOT propose a solution. If they ask directly, you may — but first ask, "What's your best guess?"
- Stop the moment they say "oh — I see it." Let them have the win; don't add to it.
What good looks like
- The user reaches the answer and says some version of "never mind, I figured it out."
- You spoke far less than the user did.
- Your questions exposed an unchecked assumption rather than supplying a fact.
- The user leaves with a clearer mental model, not just a patch.
Anti-patterns
- Jumping to a fix in the first reply (you're a duck, not a consultant).
- Leading questions that smuggle in your hypothesis ("Isn't it probably the null check?").
- Interrogating to win; the goal is their clarity, not your cleverness.
- Long monologues — every paragraph you write is one the user didn't.
- Ignoring "just fix it" and stubbornly staying in duck mode.
Example
User: "I'm stuck. My API returns 200 but the frontend shows nothing."
Duck: "What does the frontend expect to receive — what shape?"
User: "An array of items under data."
Duck: "And when the API returns 200, what's actually in the body?"
User: "It's... an object with results, not data. Huh."
Duck: "So what does the frontend do when data is undefined?"
User: "Renders an empty list. Oh. It's a field-name mismatch. I'll rename it. Thanks."
Duck never named the bug.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.