printpractical/agent-toolkit

brainstorm

Use when an idea, problem, or possible solution is not formed enough to begin architect.

First seen Jul 20, 2026

Installation

$ npx skills add printpractical/agent-toolkit --skill brainstorm

Summary

  • Use when an idea, problem, or possible solution is not formed enough to begin architect.
  • Explores the problem and alternative approaches, challenges assumptions during convergence, and can produce an optional architect-seed.md handoff for a later architect session.
  • Do NOT use once the change is ready for architectural decisions — use architect directly.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from printpractical/agent-toolkit · top by installs.

npx skills add printpractical/agent-toolkit

Browse all from printpractical/agent-toolkit

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

Stars 1
Default branch main
Open issues 0
Status Active

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 3,788 B
  • docs SUMMARY.md 375 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 37 installs

SKILL.md

Brainstorm

You are running the brainstorm entry ramp. Your job is to turn an early, incomplete idea into a clearer set of options and, when useful, a provisional direction for architect to challenge. This is optional discovery, not a pipeline phase.

Boundaries

  • Do not create a change manifest, approve an approval, or create an active change.
  • Do not present an explored option as approved architecture.
  • architect remains the right direct entry point when the user already has a sufficiently formed change.

Intake

Before reading references, exploring the repository, or creating a seed, obtain the goal and observable outcome, affected area, constraints and anti-goals, and whether requirements are formed, partially formed, or unformed.

Your stance

Read references/challenge-protocol.md. Be expansive while generating options and constructively adversarial while narrowing them. Do not dismiss an early idea for being incomplete; expose the uncertainty that prevents a sound architectural decision.

Phase 1: Frame the idea

Start with the problem, not a proposed implementation. Work one topic at a time:

  1. What problem or opportunity prompted this?
  2. Who is affected, and what outcome would make this worthwhile?
  3. What is known evidence versus a hypothesis or preference?
  4. What constraints, dependencies, and anti-goals are already clear?

If the user starts with a solution, restate the underlying problem and confirm it before evaluating the solution.

Phase 2: Diverge

Generate several meaningfully different approaches. Include retaining the status quo when it is a credible option. For each approach, capture:

  • What it changes and what it deliberately leaves alone
  • Expected benefits
  • Costs, risks, and failure modes
  • Assumptions that need validation
  • Information that would distinguish it from the alternatives

Do not force convergence before the options are genuinely distinct. Use relevant repository context or lightweight exploration when it would materially improve the comparison, but do not begin detailed architecture or implementation planning.

Phase 3: Converge

Compare the strongest options against the desired outcomes and constraints. Challenge unjustified assumptions, accidental scope expansion, and choices that create complexity without a clear payoff. State a provisional recommendation only when its tradeoffs are understood.

If the evidence is insufficient, recommend the smallest discovery step needed before architect rather than manufacturing a decision.

Phase 4: Optional handoff

Summarize the problem, options, tradeoffs, recommendation, and open questions. Ask whether the user wants a durable handoff for architect.

If yes, fill references/templates/architect-seed.md.tmpl and write architect-seed.md at the project root or another path the user chooses. Label every entry as fact, hypothesis, preference, or provisional recommendation where applicable.

Then tell the user:

"The brainstorm seed is ready. Run architect and point it at this seed. It will challenge and formalize the direction before any architect approval is approved."

If no, leave the result as session discussion. Do not create an artifact merely for bookkeeping.

Reference files

  • references/challenge-protocol.md
  • references/change-lifecycle.md
  • references/templates/architect-seed.md.tmpl