How do I install the Brainstorm skill?

Run npx skillmds@latest add ajitta/brainstorm in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Brainstorm safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Brainstorm?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Brainstorm free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Brainstorm?

ajitta (@ajitta) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Brainstorm by ajitta — Agent Skill

Name: Brainstorm
Author: ajitta

Brainstorm

Brainstorm the intervention — spread ~10 codebase-grounded candidates from ship-this-afternoon to quarter-long bet. Use on "brainstorm interventions", "show me options", or a problem named with no approach chosen. NOT for UI/design variants — use prototypes.

ajitta b872a56 2.6 KB Updated Sep 17, 2026

File contents

Brainstorm — Solution-Space Map (Brainstorm the Intervention)

Jumping on the first solution leaves the rest of the solution space unknown. The user picks; the model draws the space. Origin: Brainstorm the Intervention — see skills/loop/references/talk-source.md.

Iron Rules

Implement nothing yet. Output is a map of solutions, not a solution.
Not 10 near-duplicates — spread candidates diverse in time horizon and approach.

Procedure

Get problem definition and constraints (timeline, people, cost) from $ARGUMENTS + conversation context. If empty or vague, ask back once for a measurable statement: which metric, how bad, since when, for whom — one round only.
When a codebase is in scope, search it first (or spawn the unknowns:unknowns-scout agent with the Agent tool) for machinery that already exists but sits disconnected — the cheapest candidates are usually wiring, not building.
Generate ~10 candidate interventions, spread across the time axis: ship this afternoon / short-term (1–2 weeks) / mid-term (quarter) / quarter-long bets.
Attach to each candidate: name / one-line description / expected impact (what improves, by how much) / effort size (S/M/L/XL) / key risks / how to measure success, plus Found in code — <path> when it builds on machinery already in the repo.
Present arranged by impact × effort (a plugin extension of the source's single time axis) — quick wins and big bets distinguishable at a glance.
After user selection: convert chosen interventions into structured next steps — down to each intervention's first execution prompt (or experiment design).

Output

Artifact tool if available → else .unknowns/<YYYY-MM-DD>-brainstorm-<slug>.html → else markdown; always echo the assembled reply in chat. Cards on a time axis (ship this afternoon → quarter-long bet) with a toggle to the impact × effort matrix; each card's resonate checkbox assembles into a copyable "proceed with these, in this order". Details: skills/loop/references/output-routing.md

Follow-ups

After picking interventions: form in question → /unknowns:prototypes (or /prototypes for copied installs); spec finalization → /unknowns:interview; implementation plan → /unknowns:plan.

---name: brainstormdescription: Brainstorm the intervention — spread ~10 codebase-grounded candidates from ship-this-afternoon to quarter-long bet. Use on "brainstorm interventions", "show me options", or a problem named with no approach chosen. NOT for UI/design variants — use prototypes.---# Brainstorm — Solution-Space Map (Brainstorm the Intervention)Jumping on the first solution leaves **the rest of the solution space unknown**.The user picks; the model draws the space.Origin: Brainstorm the Intervention — see skills/loop/references/talk-source.md.## Iron Rules1. **Implement nothing yet.** Output is a map of solutions, not a solution.2. Not 10 near-duplicates — spread candidates **diverse in time horizon and approach**.## Procedure1. Get problem definition and constraints (timeline, people, cost) from `$ARGUMENTS` + conversation context. If empty or vague, ask back once for a measurable statement: which metric, how bad, since when, for whom — one round only.2. When a codebase is in scope, **search it first** (or spawn the `unknowns:unknowns-scout` agent with the Agent tool) for machinery that already exists but sits disconnected — the cheapest candidates are usually wiring, not building.3. Generate **~10** candidate interventions, spread across the time axis: ship this afternoon / short-term (1–2 weeks) / mid-term (quarter) / quarter-long bets.4. Attach to each candidate: **name / one-line description / expected impact (what improves, by how much) / effort size (S/M/L/XL) / key risks / how to measure success**, plus **`Found in code — <path>`** when it builds on machinery already in the repo.5. Present arranged by **impact × effort** (a plugin extension of the source's single time axis) — quick wins and big bets distinguishable at a glance.6. After user selection: convert chosen interventions into **structured next steps** — down to each intervention's first execution prompt (or experiment design).## OutputArtifact tool if available → else `.unknowns/<YYYY-MM-DD>-brainstorm-<slug>.html` → elsemarkdown; always echo the assembled reply in chat.Cards on a time axis (ship this afternoon → quarter-long bet) with a toggle to theimpact × effort matrix; each card's **resonate checkbox** assembles into a copyable"proceed with these, in this order".Details: skills/loop/references/output-routing.md## Follow-ups- After picking interventions: form in question → `/unknowns:prototypes` (or `/prototypes` for copied installs); spec finalization → `/unknowns:interview`; implementation plan → `/unknowns:plan`.

Frequently asked questions