Creative ideation
Skill furkangonel/cowrangler/bundled_skills/creative/creative-ideation
Autonomous terminal AI agent for workflows and feasible project procedures. Co-Worker Co-Wrangler π
npx -y skills add furkangonel/cowrangler --skill creative-ideationAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Structured brainstorming sessions using proven ideation techniques.
SKILL.md
6.6 KB, as published. Nobody here has run it
Creative Ideation SOP
When to Use
- User needs ideas for a product, feature, campaign, or story
- User says "brainstorm", "I'm stuck", "give me ideas", "what could I do with..."
- User wants to explore a problem space before committing to a solution
- User needs to break out of a creative rut
Step 1 β Frame the Challenge
Before generating ideas, nail down the challenge statement. Use How Might We (HMW) format:
"How might we [verb] [object] so that [desired outcome]?"
Examples:
- "How might we reduce onboarding friction for enterprise users so that activation takes < 5 minutes?"
- "How might we make our weekly newsletter feel like a conversation so that readers reply?"
Ask the user to confirm or refine the HMW before proceeding to ideation.
The 5 Techniques
Technique 1 β SCAMPER
Apply each lens to the subject and generate at least 2 ideas per letter:
| Letter | Question | Apply to: [subject] |
|---|---|---|
| Substitute | What can be replaced? | |
| Combine | What can be merged or bundled? | |
| Adapt | What can be borrowed from elsewhere? | |
| Modify / Magnify | What can be exaggerated or scaled? | |
| Put to other uses | What else can this do? | |
| Eliminate | What can be removed entirely? | |
| Reverse / Rearrange | What if you flip the order or approach? |
Output format:
SCAMPER β [Subject]
S: [Idea 1], [Idea 2]
C: [Idea 1], [Idea 2]
A: [Idea 1], [Idea 2]
M: [Idea 1], [Idea 2]
P: [Idea 1], [Idea 2]
E: [Idea 1], [Idea 2]
R: [Idea 1], [Idea 2]
Technique 2 β Reverse Brainstorming
Instead of solving the problem, brainstorm how to make it worse. Then flip each answer.
Steps:
- Reverse the challenge: "How might we CAUSE [problem]?"
- Generate 8β10 ways to make things terrible
- Reverse each terrible idea into a potential solution
Example:
Challenge: How might we improve user retention?
Reversed: How might we guarantee users churn immediately?
Terrible ideas β Flipped solutions:
- Make onboarding confusing β Create a guided, skippable tutorial
- Never send notifications β Send personalized, timely nudges
- Hide the valuable features β Surface key features on day 1
- Make cancellation instant β Add a "pause" option before cancel
Technique 3 β Crazy 8s
Generate 8 distinct ideas in 8 minutes. Speed over quality β defer judgment entirely.
Rules:
- Each idea gets one sentence max
- No filtering, no "but that's impossible"
- Ideas should vary widely in scale, cost, and approach
- Number them 1β8
Output format:
Crazy 8s β [Topic]
1. [Wild idea]
2. [Obvious idea]
3. [Expensive idea]
4. [Free/simple idea]
5. [Tech-heavy idea]
6. [Human/offline idea]
7. [Stolen from another industry]
8. [Opposite of what's expected]
After listing, ask the user: "Which 1β3 feel most worth exploring? I'll develop those."
Technique 4 β How Might We (HMW) Expansion
Take one problem space and generate 10+ HMW questions from different angles:
Angles to cover:
- Amplify the good
- Remove the bad
- Explore the opposite
- Question assumptions
- Change the context/user
- Combine with something else
- Change the time horizon
- Change the scale (10x bigger / 10x smaller)
Example output:
HMW questions for "improve developer onboarding":
1. HMW make the first PR feel like a win, not a chore?
2. HMW remove all environment setup friction?
3. HMW turn documentation into an interactive playground?
4. HMW onboard 10 developers simultaneously without a senior engineer?
5. HMW make the codebase self-explaining?
6. HMW compress 2-week onboarding into 2 days?
7. HMW celebrate onboarding milestones automatically?
8. HMW make failing a test feel helpful instead of discouraging?
Technique 5 β Mind Map
Build a radial idea tree starting from the core concept:
[Core Concept]
/ | \
[Theme A] [Theme B] [Theme C]
/ \ | / \
[Idea] [Idea] [Idea] [Idea] [Idea]
Steps:
- Place the core concept at center
- Identify 4β6 major themes/dimensions
- Branch 2β4 ideas from each theme
- Mark cross-connections between branches
Output: Use indented text or ASCII tree. Offer to convert to a visual if the user wants it.
Core: Subscription Product Retention
βββ Engagement
β βββ Weekly digest email with personalized stats
β βββ In-app streak / habit tracking
β βββ Community challenges
βββ Value Delivery
β βββ Monthly "wins" report showing ROI
β βββ Proactive feature discovery tips
β βββ Benchmarking against peers
βββ Friction Reduction
β βββ One-click pause instead of cancel
β βββ Billing flexibility (pause, downgrade)
β βββ Offboarding interview with save offer
βββ Relationship
βββ Dedicated CSM at $X threshold
βββ Early access to new features
βββ Annual user conference / virtual event
Facilitation Flow
When a user asks for ideation help, follow this sequence:
- Frame β Help them write a clear HMW statement (1β2 exchanges)
- Diverge β Run 1β2 techniques and generate raw ideas without filtering
- Cluster β Group similar ideas and name the clusters
- Converge β Ask which clusters/ideas to develop further
- Develop β Flesh out the top 2β3 ideas with next steps
Output Format for Final Ideas
For each selected idea, provide:
## Idea: [Name]
**One-liner:** What it is in one sentence.
**How it works:** 3β5 bullet description.
**Why it might work:** The key insight or assumption.
**Risks / unknowns:** What needs to be validated first.
**Next step:** The smallest experiment to test this.
Agent Instructions
- Start every session by confirming the challenge framing β don't generate ideas on a vague prompt
- Choose the technique that best fits the context:
- Existing product β SCAMPER or Reverse Brainstorming
- Open-ended problem β HMW Expansion or Mind Map
- Fast divergence needed β Crazy 8s
- Generate at least 8β10 raw ideas before filtering
- Never evaluate ideas during divergence β just list them
- After divergence, explicitly shift: "Now let's narrow down β which of these resonates?"
- Always end with a concrete next step for the chosen idea