brainstorm
Skill2 starsChanged 7 days agoMITView on GitHub
Ideation organ upstream of /envision: talk a half-formed idea into a name, one-line description, edges (in/out) and end-goal, and write a light .claude/ideas/<slug>.md note that /envision picks up. Converges rather than chats. Use on /brainstorm [slug], \"help me shape this idea\", or when the user has a vague idea to shape. Mode A — propose-confirm, never auto-fires.
189 other projects use a version of this.
Adopted by
Most recently taken up by
- nielsmadan/skillstheir own version, first seen Nov 8, 2026
- phuthuycoding/kanban-flowtheir own version, first seen Nov 8, 2026
- baethon/skillstheir own version, first seen Nov 8, 2026
- CloudChef/atlasclaw-providerstheir own version, first seen Nov 8, 2026
- nisarth/claude-skills-storetheir own version, first seen Nov 8, 2026
- LongJohnSilver1504/claude-skillstheir own version, first seen Nov 8, 2026
Used with
Projects with it also use playtest-report, team-polish 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.
CoralGameSkillplaytest-report
Generates a structured playtest report template or analyzes existing playtest notes into a structured format. Use this to standardize playtest feedback collection and analysis.
Also in 9 projects that use this
CoralGameSkillteam-polish
Orchestrate the polish team: coordinates performance-analyst, technical-artist, sound-designer, and qa-tester to optimize, polish, and harden a feature or area for release quality.
Also in 11 projects that use this
lilangMaxSkillsoak-test
Generate a soak test protocol for extended play sessions. Defines what to observe, measure, and log during long play sessions to surface slow leaks, fatigue effects, and edge cases that only appear after sustained play. Primarily used in Polish and Release phases.
Also in 6 projects that use this
lilangMaxSkillteam-live-ops
Orchestrate the live-ops team for post-launch content planning: coordinates live-ops-designer, economy-designer, analytics-engineer, community-manager, writer, and narrative-director to design and plan a season, event, or live content update.
Also in 6 projects that use this
- +8 more from CoralGame
Other versions189 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 189
- Donchitos/Claude-Code-Game-Studios★ 27ksame name
- codeaholicguy/ai-devkit★ 1.7ksame name
- jtydhr88/ComfyTV★ 1.2ksame name
- event-driven-io/emmett★ 562same name
- zaxbysauce/opencode-swarm★ 519same name
- ZaxbyHub/opencode-swarm★ 518same name
- automagik-dev/genie★ 351same name
- namastexlabs/automagik-genie★ 351same name
- bkywksj/knowledge-base★ 340same name
- pixel-cellar/Claude-Code-Game-Studios★ 324same name
179 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 repository47 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .github/copilot-instructions.mdCopilot instructions
- autonomizeSkill
- blueprintSkill
- captureSkill
- checkpointSkill
- close-outSkill
- coordinateSkill
- cycleSkill
- deliverSkill
- design-proposalSkill
- discoverSkill
- dispatchSkill
- draft-planSkill
- envisionSkill
- executeSkill
- extract-conventionsSkill
- glossarySkill
- humanizeSkill
- human-viewSkill
- improve-promptSkill
- memory-auditSkill
- new-agentSkill
- new-agents-mdSkill
- new-hookSkill
- new-mcpSkill
- orchestrateSkill
- pr-authorSkill
- prepSkill
- pre-pushSkill
- pr-triageSkill
- recapSkill
- retrospect-coreSkill
- reviewSkill
- rulesSkill
- shortcutsSkill
- specSkill
- systematic-debuggingSkill
- test-driven-developmentSkill
- update-changelogSkill
- update-codemapSkill
- using-git-worktreesSkill
- visualise-codemapSkill
- visualSkill
- work-stateSkill
- wrapSkill
- write-testsSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 156 edited versions we read keep these.
4
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
- Distinctive: line 21
Ask one question. If the user hasn't said why they want this, ask why, openly and in their terms, with no options and no example answers.
- Distinctive: line 28
No numbered question lists, no option menus, no draft name or note in the first reply.
- Very specific: line 41
If .claude/ideas/<slug>.md exists, read it and resume from where it left off (this is the cross-session-gap recovery the persisted note …
- Distinctive: line 45
If the idea points at something (a project, files, notes, an earlier conversation), read it before your first reply.
- Distinctive: line 62
When the five are roughly there (or the user says "pin it"), propose the crystallized note for accept / edit / reject:
- Very specific: line 81
Don't force premature precision — status: shaping is allowed if it genuinely isn't crystallized yet, but say so and name what's still …
“Common” is measured: kept by at least 6 in 10 of the 156 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
brainstorm
The ideation organ. The stage before /envision: where a half-formed or blurry idea gets talked into enough shape that /envision can take it. One organ, two moods — discuss (explore a loose idea) and crystallize (pin a blurry one down) — they're the same conversation at different stages, so there's one verb.
[half-formed idea] → /brainstorm → shaped idea (.claude/ideas/<slug>.md) → /envision → /spec → …The discipline: it must converge, not just chat. A /brainstorm that ends without a crystallized statement the user accepts was just conversation. The engineer wants the idea sharpened, not a transcript.
Your first reply
The first reply is short and has exactly this shape:
- Play the idea back in two or three sentences, in your words.
- Ask one question. If the user hasn't said why they want this, ask why, openly and in their terms, with no options and no example answers.
- Say in one line what the answer will change.
Example, for "/brainstorm a tool that tracks my reading":
You want something that keeps a record of what you read, so it doesn't vanish once you close the tab or book. What made you want this now? Knowing whether it's about remembering, finding things again or reading more decides what the tool keeps.
No numbered question lists, no option menus, no draft name or note in the first reply. If the idea is already sharp and the why is known, skip the question and propose the crystallized note instead.
When to run
/brainstorm [slug]— start (or resume) shaping an idea.- Natural language: "help me shape this idea", "I have a vague idea for X", "let's think this through", "crystallize this", "put a name and edges on this".
- Do not auto-fire. Mode A. Not for an idea that's already shaped (go straight to
/envisionor/spec), not for a one-line task (just do it).
Procedure
1. Resolve slug + read any prior note
- Slug given → use it. Omitted → derive a short kebab slug from the idea once it has a name, or ask.
- If
.claude/ideas/<slug>.mdexists, read it and resume from where it left off (this is the cross-session-gap recovery the persisted note buys).
2. Look before you ask
If the idea points at something (a project, files, notes, an earlier conversation), read it before your first reply. Never ask what it already answers, and don't reopen what the user already decided.
3. Discuss — explore, but steer toward shape
Talk the idea through. Every few turns, pull toward the five things a shaped idea needs (below). Do not drift into spec-level detail (functions, files, schemas): that's /spec's job. When the user starts there, redirect: "that's implementation — let's pin what this is first."
After that, one question per message. Ask only when the answer would change the name, the edges or the end goal and nothing already answers it; otherwise state your call as a proposal the user can correct. Once you know the why, when a choice has options you can name, sketch each with its trade-off and say which you'd pick.
The five things to converge on:
- Name — what to call it.
- Description — one paragraph: what it is, who it's for, the one-sentence success criterion.
- Edges — what's in and, more importantly, what's explicitly out (the non-goals that keep it from absorbing scope).
- End goal — what "done / working" looks like.
- Open edges — the questions still genuinely unresolved (carried forward, not forced).
4. Crystallize — propose the shaped idea
When the five are roughly there (or the user says "pin it"), propose the crystallized note for accept / edit / reject:
---
slug: <slug>
status: shaped
created: YYYY-MM-DD
description: <one-line>
---
# Idea — <name>
**What it is:** <one paragraph — what, who for, success in a sentence>
**In:** <comma-separated — what this is>
**Out (non-goals):** <comma-separated — what it deliberately isn't>
**End goal:** <what done/working looks like>
**Open edges:** <unresolved questions, or "none yet">Don't force premature precision — status: shaping is allowed if it genuinely isn't crystallized yet, but say so and name what's still blurry.
5. Write + hand off
On accept: write .claude/ideas/<slug>.md, update .claude/ideas/INDEX.md (one row per idea: slug / status / created / one-line). Then point downstream:
Shaped
<name>→.claude/ideas/<slug>.md. When you're ready to partition it into modules, run/envision <slug>— it'll pick up this note as the starting intent.
On edit → revise the draft, re-propose. On reject → discard, write nothing.
What brainstorm does NOT do
- Does not partition into modules / draw the module map. That's
/envision. brainstorm stops at the shaped idea; envision turns it into a structure. - Does not produce spec-level detail (functions, files, data shapes, APIs). Redirect to
/spec, downstream. - Does not just chat. It must converge to a crystallized note the user accepts, or it produced nothing.
- Does not auto-fire, and does not silently write — the note is proposed first.
- Does not implement anything.
Relationship to other organs
- envision — the immediate downstream. brainstorm's
.claude/ideas/<slug>.mdbecomes/envision <slug>'s starting intent. - improve-prompt — a sibling interview-to-converge organ, but for prompts, not project ideas. Different corpus, same converge-don't-ramble discipline.
- spec — two hops downstream (brainstorm → envision → spec). brainstorm shapes the idea; spec pins one module's requirements.
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.

