# Lateral Thinking

> Use lateral thinking to break out of linear reasoning and near-duplicate ideas by changing perspective, reversing assumptions, and exploring unexpected angles. Use when current thinking is stuck, options are too similar, or the team needs a non-obvious angle rather than more of the same logic.

- Skill: `alubiama/lateral-thinking` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add alubiama/lateral-thinking`
- Raw SKILL.md: https://api.skillmd.com/api/skills/alubiama/lateral-thinking/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: Alubiama (https://skillmd.com/u/alubiama)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/alubiama/lateral-thinking

---


# Lateral Thinking

## Goal

Break a stuck line of reasoning and produce a few non-obvious directions worth testing.

The job of this skill is not to be weird for its own sake. The job is to interrupt stale logic, challenge hidden assumptions, and create useful directional shifts.

This skill opens new angles. It does not claim those angles are validated or superior unless evidence is explicitly provided.

## Default Posture

- disrupt the frame, not reality
- challenge assumptions before adding features
- seek useful surprise, not randomness
- produce a few sharp perspective shifts
- reconnect novelty back to the real problem

## When To Use

Run this skill when:

- current ideas are near-duplicates
- reasoning is too linear or too constrained by precedent
- the team keeps refining the same weak concept
- a fresh angle is needed before deeper selection
- conventional ideation is producing only incremental moves

This is a strong next step after:

- `scamper` when variants remain too similar
- `jtbd` when the job is clear but thinking is still trapped
- before `creative-problem-solving` when fresh reframing is needed

## Scope Boundaries

In scope:

- challenge assumptions and defaults
- shift vantage points and reframing moves
- produce a small set of unexpected but relevant angles
- reconnect surprising directions to practical next tests

Out of scope by default:

- random idea generation with no relevance
- replacing root-cause framing
- solving formal contradictions better handled by `triz`
- treating novelty as evidence of value

## Escalation Conditions

Pause and keep uncertainty explicit when:

- the current frame cannot be stated clearly
- there is no dominant assumption to break
- the output becomes gimmicky rather than useful
- the real need is framing, contradiction solving, or validation instead

If angle quality remains weak after one pass, recommend returning to `jtbd`, `scamper`, `triz`, or method selection.

## Lateral Workflow

1. Restate the current problem frame.
2. Name the dominant assumptions or default logic.
3. Pick 2-4 lateral moves that best disrupt the stuck frame:
   - assumption reversal
   - inversion
   - random entry
   - forced analogy
   - perspective shift
   - constraint removal
4. Generate one useful angle from each chosen move.
5. Remove ideas that are merely strange, not helpful.
6. Translate the strongest angles back into actionable next directions.
7. Recommend the next comparison, experiment, or downstream method.

## Lateral Moves

Prefer:

- `assumption reversal` when hidden defaults are driving the solution space
- `inversion` when the standard order or logic may be backwards
- `random entry` when the space is overly predictable
- `forced analogy` when another domain may reveal structure the current one hides
- `perspective shift` when another actor sees a different problem
- `constraint removal` when an assumed limit may be optional

## Core Questions

Use the minimum set needed:

- What assumption is currently shaping all ideas?
- What if the opposite were true?
- What other actor would frame this differently?
- What unrelated domain solved a structurally similar problem?
- Which surprising angle is still useful after scrutiny?
- What is the smallest way to test this reframing?

## Output Contract

Always return:

1. `current_frame`
2. `stuck_pattern`
3. `assumptions_to_break`
4. `selected_moves`
5. `new_angles`
6. `discarded_angles`
7. `evidence_status` (`provided`, `inferred`, `mixed`)
8. `open_unknowns`
9. `recommended_followup` (`lateral-thinking`, `jtbd`, `scamper`, `triz`, `selector`, `validation`)
10. `next_test`

## Good Lateral Signals

Strong signals:

- current options all sound alike
- a dominant assumption is easy to name
- the team is overfitting to precedent
- a perspective shift could materially change direction

Weak signals:

- the problem is still undefined
- the issue is really a contradiction
- the baseline is missing and ordinary variation has not been tried
- the task needs disciplined evaluation, not more divergence

## Guardrails

- Do not confuse novelty with usefulness.
- Do not use lateral thinking to avoid defining the real problem.
- Do not replace `triz` when a true contradiction exists.
- Distinguish evidence from inference.
- If angles remain gimmicky or irrelevant, recommend returning to `jtbd`, `scamper`, or method selection.
- Prefer a few useful reframings over a pile of strange prompts.
- If no meaningful frame break appears, say so instead of padding the list.

## Anti-Patterns

Watch for these failure modes:

- random nonsense presented as creativity
- perspective shifts with no link back to the job
- breaking assumptions that were actually valid constraints
- generating many quirky angles and no actionable next move
- using lateral thinking as a substitute for framing or validation

## Example

User request:

`Every idea for our onboarding still feels like the same checklist with different wording.`

Expected shape of response:

1. `current_frame`: onboarding is a checklist users must complete before receiving value
2. `stuck_pattern`: all ideas optimize the same checklist model
3. `assumptions_to_break`: value must come after onboarding; every user needs the same path; education must be front-loaded
4. `selected_moves`: `assumption reversal`, `perspective_shift`, `inversion`
5. `new_angles`: let users start before full onboarding; trigger guidance only after uncertainty signals; design onboarding from the impatient user's perspective
6. `discarded_angles`: abstract "gamify it" direction with no changed mechanism
7. `evidence_status`: `inferred`
8. `open_unknowns`: where users actually stall, which tasks create confidence, which segments need guidance
9. `recommended_followup`: `validation`
10. `next_test`: compare a start-before-teaching concept with the current checklist flow