agentleFS
fightZyfightZy/simple-skills/.agents/skills/advanced-brainstorming/SKILL.md

advanced-brainstorming

Skill1 starsChanged 12 days agoMITView on GitHub

Use when a user wants idea generation to become broader, less conservative, more imaginative, or more frame-breaking, especially when expanding an existing proposal into multiple unconventional but still intelligible directions.

The file

Show its rules and text

We split this skill into the rules it gives an agent

22 rules. About 0 are common, 22 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈22

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this skill when the user wants genuine divergence, not the safest answer with nicer wording.

    line 8
  2. Distinctive:

    Do not treat the user's first framing as the problem statement.

    line 35
  3. Distinctive:

    Follow this sequence unless the user explicitly asks for a shorter format.

    line 45
  4. Distinctive:

    Before generating alternatives, name the default assumptions you see in the proposal.

    line 49
  5. Distinctive:

    Look for assumptions about:

    line 51
  6. Distinctive:

    Do not spend a long time explaining them.

    line 62
  7. Distinctive:

    Do not immediately replace the original idea.

    line 68
  8. Distinctive:

    After extending the original direction, deliberately challenge it.

    line 72
  9. Distinctive:

    Import patterns from outside the obvious category.

    line 87
  10. Distinctive:

    Do not name operators mechanically unless it helps the user.

    line 98
  11. Distinctive:

    Do not dump one flat list.

    line 102
  12. Distinctive:

    If the user asks for next steps after this, that is a separate phase.

    line 118
  13. Distinctive:

    Default to this structure:

    line 122
  14. Distinctive:

    If the change is not clear, the idea is probably fake novelty.

    line 141
  15. Distinctive:

    do not return only one "best" idea

    line 145
  16. Distinctive:

    do not drift into generic feature lists

    line 146
  17. Distinctive:

    do not confuse novelty with randomness

    line 147
  18. Distinctive:

    do not reward safety just because it is easier to explain

    line 148
  19. Distinctive:

    do not collapse into execution steps unless the user explicitly asks

    line 149
  20. Distinctive:

    keep at least one section that meaningfully breaks the original frame

    line 150
  21. Distinctive:

    keep at least one section that stretches farther than you would normally be comfortable with

    line 151
  22. Distinctive:

    If you notice these thoughts, widen the search space before finishing:

    line 165

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Advanced Brainstorming

Use this skill when the user wants genuine divergence, not the safest answer with nicer wording.

The job is to:

  1. widen the possibility space before converging
  2. challenge the assumptions inside the user's framing
  3. generate multiple unconventional directions that are still legible
  4. end with light prioritization instead of execution planning

This skill is for expansion, not validation, research, or implementation.

Use This Skill For

  • expanding an existing proposal into stronger and less obvious directions
  • opening up a vague concept without collapsing into generic idea lists
  • helping the user escape conservative, polished, low-novelty answers
  • generating multiple routes, including frame-breaking and wildcard routes

Do Not Use This Skill For

  • market validation or competitor research
  • deciding whether an idea should continue, pivot, or stop
  • detailed product plans, task breakdowns, or implementation steps
  • coding, debugging, or execution work

Core Principle

Do not treat the user's first framing as the problem statement. Treat it as the first doorway into the possibility space.

Good advanced brainstorming does three things at once:

  • keeps enough contact with the user's intent to stay useful
  • deliberately changes important axes instead of only decorating the same idea
  • preserves multiple strong possibilities before lightly converging

Workflow

Follow this sequence unless the user explicitly asks for a shorter format.

1. Surface the hidden assumptions

Before generating alternatives, name the default assumptions you see in the proposal.

Look for assumptions about:

  • target user
  • core problem
  • product shape
  • distribution channel
  • business model
  • interface or interaction style
  • time horizon
  • what the user is optimizing for

Do not spend a long time explaining them. State them cleanly and move on.

2. Expand along the current vector

First, make the user's current direction stronger, sharper, or more differentiated.

Do not immediately replace the original idea. Show that you understand its momentum, then push it further than the obvious MVP path.

3. Break the frame

After extending the original direction, deliberately challenge it.

Change one or more major axes:

  • a different user
  • a different problem behind the stated problem
  • a different delivery mechanism
  • a different business model
  • a different social unit such as individual, team, or network
  • a different time scale such as real-time, episodic, or long-term

The user's proposal is one candidate frame, not a contract.

4. Cross-pollinate

Import patterns from outside the obvious category.

Useful idea operators:

  • inversion: what if the product does the opposite of the default pattern
  • role reversal: what if the user cares for the system rather than the reverse
  • environment shift: what if the product is a place, ritual, or atmosphere rather than a tool
  • time shift: what if the value appears over months instead of in-session
  • audience shift: what if the real buyer or beneficiary is different
  • combination: what two adjacent ideas create a stronger hybrid

Do not name operators mechanically unless it helps the user. Use them to create better directions.

5. Split ideas into layers

Do not dump one flat list.

Organize the output into layers such as:

  • practical unconventional directions
  • bigger bets with higher upside and more risk
  • wildcards that may sound strange but unlock better thinking

This protects breadth without turning the whole answer into noise.

6. Lightly converge

End by selecting the two or three most promising directions.

Explain why they are promising, but stop before turning the response into an implementation plan.

If the user asks for next steps after this, that is a separate phase.

Output Contract

Default to this structure:

  1. Hidden Assumptions
  2. Expanded Directions
  3. Frame-Breaking Alternatives
  4. Wildcards
  5. Most Worth Exploring Next

Each direction should clearly state what changed.

Good changes include:

  • the user changed
  • the job-to-be-done changed
  • the product shape changed
  • the interaction model changed
  • the business model changed
  • the time horizon changed

If the change is not clear, the idea is probably fake novelty.

Quality Bar

  • do not return only one "best" idea
  • do not drift into generic feature lists
  • do not confuse novelty with randomness
  • do not reward safety just because it is easier to explain
  • do not collapse into execution steps unless the user explicitly asks
  • keep at least one section that meaningfully breaks the original frame
  • keep at least one section that stretches farther than you would normally be comfortable with

Common Failure Modes

Failure modeCorrection
Safe incrementalismChange a meaningful axis, not just feature depth
Generic brainstorm dumpUse the required output sections and layers
Fake boldnessState exactly what changed from the original framing
Wild but useless ideasKeep each direction interpretable and tied to the user's intent
Premature execution adviceStop at light prioritization

Red Flags

If you notice these thoughts, widen the search space before finishing:

  • "I already know the best answer"
  • "the original framing should probably stay fixed"
  • "these stranger options are too risky to include"
  • "I should jump to MVP or execution guidance now"
  • "this is enough variety" when all options still share the same core shape

Micro-Example

User proposal:

Build an AI bookkeeping tool for indie developers.

Weak response:

  • add auto-categorization
  • add tax reports
  • add Stripe sync

Stronger response shape:

  • Hidden assumption: the product must stay a bookkeeping tool
  • Expanded direction: AI CFO for solo builders
  • Frame-breaking alternative: revenue operating system across multiple products
  • Wildcard: benchmark network for anonymous solo-business finance patterns
  • Most worth exploring next: AI CFO plus benchmark layer

The goal is not to be louder. The goal is to expand the conceptual space in ways the user can still reason about.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.