grill-me
Skill0 starsChanged 4 months agoMITView on GitHub
Use PROACTIVELY whenever Gera is working something OUT OF HIS HEAD rather than asking a quick question — he's mulling or scoping a plan/design/decision, describing a process or how he does something, floating a half-formed idea, or wants to capture/document/scope knowledge before building. Natural signals (he will rarely say "grill me"): "I've been thinking about…", "I have an idea for…", "I want to build/plan/figure out…", "I'm not sure how to approach…", "let me think this through", "help me scope/organize this", "here's how I do X", "I keep doing X by hand", or any long mulling-out-loud message. Also the explicit "grill me", "interrogate me", "stress-test this", "extract this into the wiki", "/grill-me". The relentless one-question-at-a-time extraction engine that captures his thinking to a durable doc and graduates it into the wiki. SAFETY: unless he explicitly asked to be grilled, OFFER first ("want me to grill you on this and capture it?") and wait for yes — never silently launch a long interview on a quick question. Distinct from brainstorming (explores NEW design space) and the builders' discovery interviews (tied to building ONE artifact); grill-me extracts EXISTING knowledge/decisions Gera already holds.
193 other projects use a version of this, 10 of them edited.
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 · 10
- openstatusHQ/openstatus★ 9.3k10% same
- sanity-io/sanity★ 6.4k10% same
- camunda/camunda★ 4.4k10% same
- openstatusHQ/data-table-filters★ 2.3k10% same
- nateherkai/AIS-OS★ 1.4k22% same
- jewbetcha/openflight★ 1.2k10% same
- truffle-ai/dexto★ 65910% same
- brianlovin/claude-config★ 37610% same
- kubev2v/forklift★ 21310% same
- mart337i/odoo-skills★ 4311% same
Wrote their own version · 183
- 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
- webiny/webiny-js★ 8ksame name
173 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 repository50 other files it gives its agents.
- aegisSkill
- agent-builderSkill
- agents-team-builderSkill
- ai-applySkill
- auditSkill
- bottleneckSkill
- buildSkill
- capability-radarSkill
- clay-leadsSkill
- cowatchSkill
- deepforkSkill
- design-blueprintSkill
- evalSkill
- fable-modeSkill
- financesSkill
- forgeSkill
- framework-builderSkill
- hooks-builderSkill
- level-upSkill
- loops-builderSkill
- meeting-hudSkill
- multi-brainSkill
- nate-watchSkill
- onboardSkill
- plugin-builderSkill
- ponytailSkill
- presentSkill
- prompt-evalSkill
- radar-harvestSkill
- rag-forgeSkill
- rlsgateSkill
- rnd-loopSkill
- rndSkill
- roastSkill
- routines-builderSkill
- servy-routineSkill
- session-handoffSkill
- skill-builderSkill
- skill-harvestSkill
- sparkSkill
- storm-researchSkill
- studySkill
- talkSkill
- taste-skillSkill
- verify-loopSkill
- verify-specificsSkill
- wheel-of-lifeSkill
- workanaSkill
- workaround-builderSkill
- workflow-builderSkill
The file
We split this skill into the rules it gives an agent
11 rules. 0 are common, 9 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 177 edited versions we read keep these.
9
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Don't collide with neighbors: if he's exploring brand-new design space, that's brainstorming.
- Distinctive: line 41
After EVERY answer Gera gives, BEFORE you ask the next question:
- Distinctive: line 43
Update or correct earlier entries: if a later answer changes them.
- Distinctive: line 44
Only then: ask the next question.
- Distinctive: line 46
Never batch multiple answers into one write.
- Distinctive: line 50
Ask one question at a time.
- Distinctive: line 52
Explore instead of asking: If a question can be answered by reading a file, the codebase, or the wiki, do that instead of asking.
- Distinctive: line 53
Flag and move on: When Gera can't answer something (only a stakeholder knows — e.g.
- Distinctive: line 79
Final reconciliation pass: read the whole capture file for contradictions or gaps and fix them.
- Very specific: line 88
If Gera defers graduation, the capture stays in knowledge/grills/ and you log a one-line follow-up in pending.md.
- Distinctive: line 98
Don't invent answers: If only a stakeholder knows, flag it with an owner — don't fill the gap with a plausible guess.
“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
What this skill does
Relentlessly interview Gera about every aspect of a topic until you reach shared understanding. Walk down each branch of the decision tree, resolving dependencies one by one. The real goal is to extract what's in Gera's head into a durable, organized markdown file so nothing is lost as context fills up — and then graduate that capture into Servy's wiki so the knowledge compounds.
This is the general-purpose extraction engine Servy was missing. The builders (/skill-builder, /agent-builder, …) each run a Discovery Interview tied to building one artifact. This skill grills Gera about anything — a business process, a Jumpstart engagement detail, a plan, a design, a codebase decision — for its own sake, and files the result.
Source + doctrine: references/grill-me-skill.md. Original by Matt Pocock; Nate Herk's checkpointing fork in knowledge/grill-me-source/grill-me-SKILL.md. Servy adaptation: captures live in knowledge/grills/, not a parallel brainstorms/ folder, and they graduate into context/ or references/ per references/wiki-protocol.md.
When to fire — and offer vs. just go
This skill should catch Gera even when he never says "grill me." Watch for the situation, not the phrase: he's mulling a plan/design/decision out loud, describing a process or "how I do X," floating a half-formed idea, or scoping something before a build. That's a grill opportunity.
But the cost of getting it wrong is asymmetric — silently launching a 30-question interview when he wanted a quick answer is worse than a missed trigger. So:
- He explicitly asked ("grill me", "interrogate me on this", "/grill-me", "stress-test this") → just go. Start the interview.
- You inferred the opportunity (he didn't say it) → OFFER first, one line, then wait: "Sounds like there's a lot in your head here — want me to grill you on it and capture it into the wiki? (it's a relentless one-question-at-a-time session)" Only start if he says yes. If he says no or just wants the quick answer, drop it and answer normally.
- It's a genuine quick question (a fact, a small fix, a yes/no) → do not fire. No offer.
Don't collide with neighbors: if he's exploring brand-new design space, that's brainstorming; if he's building one specific artifact, that's the relevant builder's own discovery interview. Grill-me is for extracting knowledge and decisions he already holds.
The capture file is the whole point
Long interviews fill up context. If you hold answers only in your head, you will eventually misremember, conflate, or drop something. So you checkpoint to disk after every single answer. The file, not your context, is the source of truth. Never make Gera ask you to save progress.
Setup — do this BEFORE the first question
- Pick the topic + slug. From
$ARGUMENTSif given, else ask Gera one line: "What are we grilling on?" Make a kebab slug. - Get today's date:
date +%F. - Create the capture file at
knowledge/grills/{YYYY-MM-DD}-{topic-slug}.md(create theknowledge/grills/folder if it doesn't exist). Every grill capture lives here — one predictable home, regardless of topic. Do NOT scatter captures into project folders or invent abrainstorms/folder. - Seed it immediately with the template below: title, date, the goal of the session, empty Summary, empty Q&A log, empty Open flags.
- Tell Gera where you're saving, in one line. Then ask Q1.
The checkpoint rule (non-negotiable)
After EVERY answer Gera gives, BEFORE you ask the next question:
- Append a structured entry to the capture file (use Edit): the question topic, the key facts and decisions from his answer (in his words where the wording matters), and any flags (things he couldn't answer + who can).
- Update or correct earlier entries if a later answer changes them. Keep the running Summary current.
- Only then ask the next question.
Never batch multiple answers into one write. Checkpoint one answer at a time. The point is that if context is lost at any moment, the file already holds everything said so far.
Interview method
- Ask one question at a time. For each, provide your recommended answer (your best inference from his context — CLAUDE.md, the wiki, the repo) so Gera can simply confirm, correct, or redirect.
- Resolve dependencies in order: settle the upstream decision before the ones that depend on it.
- Explore instead of asking. If a question can be answered by reading a file, the codebase, or the wiki, do that instead of asking. If Gera hands you a doc, read it and only surface what's net-new.
- Flag and move on. When Gera can't answer something (only a stakeholder knows — e.g. the manager, a client, a coworker), capture it as a flag with the right owner and move on. Don't stall.
- Keep going until Gera says you're done or you've covered every branch. Near the end, offer a completeness backstop: "Anything we haven't touched?"
- Match the voice rules in
references/voice.md— warm, plain, low jargon. This is a conversation, not an interrogation transcript.
Capture file template
# {Topic}: Grill / Discovery Notes
Date: {date} · Goal: {one line}
## Summary / key decisions
(running synthesis, updated as you go)
## Q&A log
### Q1 — {topic}
- Asked: {question}
- Captured: {facts, decisions, in Gera's words where it matters}
- Flags: {open item -> owner}
...
## Open flags (pending input)
- {item} -> {who can answer}At the end — reconcile, then graduate (the Servy step)
- Final reconciliation pass: read the whole capture file for contradictions or gaps and fix them. Make the Summary stand on its own.
- Recap to Gera in chat: what's captured, what's still flagged (with owners), and the suggested next step.
- Offer to graduate the polished knowledge into Servy's wiki, and propose the destination per
references/wiki-protocol.md:- Business / domain / "about Gera or the work" facts → a
context/file. - Reusable knowledge, a framework, an entity/concept page → a
references/wiki page (new or folded into an existing one); then updatereferences/index.mdand append a## [date] ingestline toreferences/log.md. - A plan/spec/design that will drive a build → the relevant
docs/orprojects/spot. Name the target you recommend and why; let Gera confirm, redirect, or defer. Don't auto-write into the wiki without his yes (Bike Method — keep him in the loop).
- Business / domain / "about Gera or the work" facts → a
- Propagation scan (related artifacts). A grill usually surfaces nuance that existing artifacts are missing. After reconciliation, scan
.claude/skills/*/SKILL.mdnames + descriptions andreferences/index.mdfor anything related to the grill topic — a skill, a wiki page, a guide. If you find matches, offer Nate-style: "You have a packaging skill and a packaging page, and we covered nuance that's in neither — want me to fold it in?" On his yes, fold only the specific new nuance into each artifact (smallest faithful edit, never a rewrite). Same consent rule — propose, don't auto-write. (Nate's end-of-grill move, 2026-07-11 course [4:27:46].) - Glossary harvest (English-term capture). Grills are full of English technical terms Gera stumbled on — those belong in the working glossary so they don't evaporate. After reconciliation, run
python3 scripts/glossary-harvest.py --capture knowledge/grills/{file}.md. It proposes new terms only (bolded terms, acronyms, "what is X" asks, multiword domain phrases), already deduped againstreferences/tech-comm-glossary.md. Show Gera the proposed block, fill the<plain definition>and<say-it-in-a-meeting sentence>for the ones he keeps, and on his yes append them under the right## Categoryheading in the glossary. Same consent rule as the rest of graduation — propose, don't auto-write. Generic knowledge only; never put customer specifics on that page. - If Gera defers graduation, the capture stays in
knowledge/grills/and you log a one-line follow-up inpending.md.
Re-running
Grills are re-runnable. To deepen or update an existing capture, open the same knowledge/grills/ file and say "grill me again on this" — append new Q&A, reconcile against earlier entries, re-offer graduation.
Notes / guardrails
- One question at a time. Always. Batching defeats the purpose and overwhelms Gera.
- Never skip a checkpoint. The file is the source of truth, not your context window.
- Don't invent answers. If only a stakeholder knows, flag it with an owner — don't fill the gap with a plausible guess.
- No parallel knowledge store. Captures go in
knowledge/grills/; polished knowledge graduates into the existingcontext//references/structure. Never create abrainstorms/folder. - Don't graduate without consent. The wiki is Servy's source of truth; a grill capture is raw working material until Gera approves the promotion.
- This skill is interactive and conversational — it must stay in the live session (do not fork to an isolated agent).
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.

