grill-me
Skill0 starsChanged 4 months agoMITView on GitHub
Interview yourself relentlessly about a plan, design, or idea, one question at a time, checkpointing every answer to a running file so nothing is lost. Use when you want to stress-test a plan before committing to it, extract scattered thinking into a document, surface dependencies you have been avoiding, or run a structured discovery session on your own work. The method works on any plan, proposal, design, or decision where you are the source of the knowledge and want to pressure-test it before presenting or acting on it.
193 other projects use a version of this.
Adopted by
Most recently taken up by
- xiaoxiaosrm/dsh-mattpocock-skillstheir own version, first seen Nov 8, 2026
- ishandutta2007/Awesome-Claude-Skillstheir own version, first seen Nov 8, 2026
- microsoft/frontier-agentic-modernization-rvastheir own version, first seen Nov 8, 2026
- The-Vibe-Company/quivrtheir own version, first seen Nov 8, 2026
- anotherben/claude-harnesstheir own version, first seen Nov 8, 2026
- codedb-io/codedb-lenstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use brief-to-tasks, writing-beats and 10 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.
julianoczkowskiSkillbrief-to-tasks
Break a design brief into an ordered checklist of independently buildable tasks using vertical slices. Saves as a markdown checklist. Use when user wants to break down work, create tasks from a brief, plan implementation order, or mentions "tasks" or "breakdown".
Also in 7 projects that use this
LesaniSkillwriting-beats
Shape an article as a journey of beats, choose-your-own-adventure style. The user picks a starting beat from the raw material, you write only that beat, then offer options for where to pivot next, beat by beat, until the article reaches a natural end. Use when the user has raw material and wants to assemble it as a narrative rather than an argument.
Also in 66 projects that use this
simo97Skillrequest-refactor-plan
Use this skill when user wants to create a refactor plan.
Also in 43 projects that use this
simo97Skillmigrate-to-shoehorn
Migrate test files from `as` type assertions to @total-typescript/shoehorn. Use when user mentions shoehorn, wants to replace `as` in tests, or needs partial test data.
Also in 71 projects that use this
LesaniSkillwriting-shape
Take a markdown file of raw material and shape it into an article through a conversational session — drafting candidate openings, growing the piece paragraph by paragraph, arguing about format (lists, tables, callouts, quotes) at each step. Use when the user has a pile of notes, fragments, or a rough draft and wants help turning it into something publishable.
Also in 65 projects that use this
liufeng1976Skillsetup-ts-deep-modules
Wire dependency-cruiser into a TypeScript repo so each package is a deep module — implementation hidden in subfolders, reachable only through its entry-point files. User-invoked.
Also in 42 projects that use this
- +2 more from simo97
Other versions193 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 193
- mattpocock/skills★ 309ksame name
- QuivrHQ/quivr★ 40ksame name
- The-Vibe-Company/quivr★ 40ksame name
- langfuse/langfuse★ 37ksame name
- nrwl/nx★ 29ksame name
- windmill-labs/windmill★ 18ksame name
- GreyDGL/PentestGPT★ 16ksame name
- humanlayer/humanlayer★ 13ksame name
- jaredpalmer/kev★ 9.8ksame name
- openstatusHQ/openstatus★ 9.3ksame name
183 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 repository19 other files it gives its agents.
- construct-validity-screen/CLAUDE.mdCLAUDE.md
- role-bleed-probe/CLAUDE.mdCLAUDE.md
- adversarial-hardenSkill
- ai-proofSkill
- construct-validity-screenSkill
- divergent-solveSkill
- forgeSkill
- gate-canariesSkill
- heavy-task-plannerSkill
- overnightSkill
- parallel-researchSkill
- premise-gateSkill
- prompt-architectSkill
- role-bleed-probeSkill
- scope-firstSkill
- session-handoffSkill
- stakes-gateSkill
- voice-profileSkill
- witnessSkill
The file
We split this skill into the rules it gives an agent
7 rules. 0 are common, 6 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 177 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 20
Every raw capture lives in brainstorms/.
- Distinctive: line 26
If a question can be answered by reading a file or document the user has provided, read it first and only ask about what is genuinely …
- Distinctive: line 30
Before asking the next question, append one structured entry to the capture file:
- Distinctive: line 41
Update earlier entries if a later answer corrects or refines something already captured.
- Distinctive: line 49
When the main branches feel covered, ask: "Is there anything we have not touched that you want to capture?" Then do a final pass.
- Distinctive: line 53
Read the whole capture file.
- Distinctive: line 57
Ask one question at a time with a best-guess answer offered, write every answer to a file immediately after you receive it, surface …
“Common” is measured: kept by at least 6 in 10 of the 177 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
Grill Me
A method for extracting what is in your head and stress-testing a plan before you commit to it. You know more than you realize, and most of the gaps are things you have been quietly avoiding. This skill surfaces them one question at a time, and writes down every answer so nothing is lost.
When this earns its cost
Use it when you are about to commit to a plan but have not yet had to defend it. Use it when you keep adding detail to a plan but have not tested the seams between parts. Use it when a previous plan fell apart in a meeting because a question exposed a dependency you had glossed over. For routine tasks or things you have already thought through thoroughly, skip it.
Method
0. Set up the capture file before the first question
Create a markdown file at brainstorms/YYYY-MM-DD-{topic-slug}.md (create the brainstorms/ folder if it does not exist). Give the file a header: the topic, the date, the goal of the session, and an empty section called "Open flags." Tell the user where the file is. Then ask the first question.
Every raw capture lives in brainstorms/. If the session produces a polished deliverable later, that artifact can move to a relevant project folder, but the raw capture stays where it is.
1. Ask one question at a time
For each question, offer your best-guess answer based on what you know so far, so the user can confirm, correct, or redirect rather than having to produce an answer from nothing. Ask only one question per turn. Resolve dependencies in order: settle the upstream decision before asking about things that depend on it.
If a question can be answered by reading a file or document the user has provided, read it first and only ask about what is genuinely unclear. If the user cannot answer something, flag it and move on. Do not stall on unanswered items.
2. Checkpoint after every answer, without exception
Before asking the next question, append one structured entry to the capture file:
### Q{N} — {topic}
- Asked: {the question}
- Captured: {key facts and decisions, in the user's words where the wording matters}
- Flags: {open item -> who can answer, if any}One entry per answer. Never batch multiple answers into one write. Never hold an answer in context and write it later. The file, not your context, is the source of truth. If the session were cut off after any answer, the file should already hold everything said up to that point.
Update earlier entries if a later answer corrects or refines something already captured.
3. Keep a running summary
Maintain a "Summary / key decisions" section at the top of the file. Update it as the session progresses so anyone reading the file can get the shape of the plan without reading every Q&A entry.
4. Offer a completeness backstop near the end
When the main branches feel covered, ask: "Is there anything we have not touched that you want to capture?" Then do a final pass.
5. Close with a reconciliation
Read the whole capture file. Flag any answers that contradict each other. Summarize what is resolved, what is still open, and what the suggested next step is.
One-line version
Ask one question at a time with a best-guess answer offered, write every answer to a file immediately after you receive it, surface contradictions at the end, and do not stop until every branch is covered or the user calls it done.
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.

