← Plugin catalog
Productivity

MightShape

Grant Holt v1.0.1

Shape the Right Thing

Create a free account to save this plugin.

From the developer

Declared capabilities

  • Research
  • Analysis
  • Write

Not specified in the available declarations: limitations, pricing & account requirements.

Onboarding example

Challenge this product idea before we build it, then identify the cheapest useful experiment.
Sources & dates

Developer declarations; not independently tested. Listing last checked Nov 7, 2026. Package observed Nov 6, 2026.

View package source →

Description

From the marketplace listing

Think wider. Frame better. Build what matters. MightShape preserves the momentum of a promising idea while testing whether it solves the right human problem. It combines an adaptive human-centered design practice, ten persistent fictional collaborators with complete lives and bounded expertise, sealed independent divergence, strict evidence provenance, research-grounded synthetic inquiry, real-human interview workflows, visual affinity and process maps, rapid prototypes, and an advisory Build Gate. Optional separately deployed adapters bring facilitated team exercises to Slack, Discord, and Microsoft Teams.

Language: English · Automatically detected from descriptions.

Matches for “write code”

Exact text from the indicated source. A mention alone does not establish support for your task.

Publisher capabilities · listing

Research Analysis Write Interactive

Publisher description

Human-centered Design Thinking for Codex and ChatGPT with a deeply modeled ten-person Council, evidence-safe inquiry, visible workshop steps, and visual design artifacts.

What changed

Keywords

Files & skills

File archives

Plugin package132 files · 1.92 MBBrowse files →
design-think111 files · 337 KBBrowse files →
Skill instructions
design-think21.6 KB

View saved version →

---
name: design-think
description: Facilitate human-centered product, service, AI, workflow, policy, and systems design through iterative design thinking, a persistent ten-person fictional Council, sealed divergence, evidence provenance, Inquiry Lab, participatory workshops, visual maps, learning prototypes, and an advisory Build Gate. Use for ambiguous or solution-first ideas; Ask, Meet, or Challenge the Council; sealed rounds or Minority Reports; reframing; assumptions, evidence, or debt; synthetic or human inquiry; Reality Packets or Checks; interviews; POVs/HMWs; brainstorming; affinity clustering; human/service workflow mapping; visible work; experiments; or Slack, Discord, and Teams workshops. Do not invoke solely for a named bug, specified implementation, technical clustering or process inspection, or a bounded reversible technical spike with an explicit metric and timebox unless explicitly invoked.
---

# ◇ MightShape

Think wider. Frame better. Build what matters.

Facilitate expert human-centered design without turning it into ceremony. Treat Intake as MightShape orchestration and Empathize, Define, Ideate, Prototype, and Test as iterative modes, never a waterfall.

## Route with the smallest useful intervention

1. Execute a fully specified, reversible implementation normally. Also execute a bounded, low-consequence technical validation spike directly when its uncertainty, metric, and timebox are explicit; use one compact measurement contract rather than MightShape ceremony. Use Intake for solution-first, unfamiliar, consequential, behavior-dependent, or high-lock-in work; preserve momentum and ask at most one excellent story-first question when needed.
2. Infer the user's current **starting point** and work from there. Ask one plain-language orientation question only when explicit invocation supplies too little context to act: `Where are you today: an early hunch, some real-world evidence, a framed challenge, several concepts, a prototype, or something already live? “Not sure” is fine.` Starting point never forces earlier modes.
3. Choose `QUICK_LOOK`, `SPRINT`, `STANDARD` (default), or `DEEP_DESIGN`; choose `COMPACT`, `VISIBLE` (default for substantial work), or `WORKSHOP` when the user asks to see work along the way.
4. Identify proposed solution, user, problem, outcome, evidence, assumptions, unknowns, constraints, reversibility, cost of error, and solution lock-in. For AI, also examine necessity, augmentation versus automation, override, trust, recovery, data/privacy, model failure, fallback, and value without AI.
5. Load the minimum request-specific resources below. Do not preload the whole system. For a self-contained one-turn Intake, reframe, brainstorm, compact map, or Prototype Card with sufficient inputs, use this embedded contract without a reference read; loading a file is not evidence of rigor.
6. For sustained `STANDARD` or `DEEP_DESIGN` work, resume `.design-council/project.json` or initialize it with `python3 <skill>/scripts/dc.py init --project-root <root> --name <name> --prompt <prompt>`. Skip state for transient Quick Looks.

Core invariants: understand before defining; define before solving; diverge before converging; build to learn; test assumptions; iterate when evidence changes. Keep confidence separate from evidence strength. Never turn synthetic or Council output into human evidence. Preserve dissent and history. Keep needs solution-independent. Prototype uncertainty at the lowest useful fidelity. Test to learn. Keep the Build Gate advisory; when the user says “build it anyway,” record unresolved assumptions and debt, then proceed reversibly.

Optimize for decision quality and learning value first. Token use is a secondary diagnostic, not a reason to omit a frame, outlier, contradiction, or experiment branch that could materially change the decision. Remove repetition and decorative ceremony before reducing substantive breadth or rigor.

## Load resources progressively

Choose one primary 

Preview of the first 4,000 characters. Read full instructions →

Referenced files: 110 — file references are metadata, not package contents.

Package details

Publisher declarations from the archived package. These are separate from our research and the live service's terms.

Package license
MIT
Package author
Grant Holt
Keywords
See keywords

Declared capabilities

  • Research
  • Analysis
  • Write
  • Interactive

Package observed Nov 6, 2026.

Technical details
First seen
Sep 30, 2026 · 22:02 UTC
Last seen
Nov 7, 2026 · 00:00 UTC
Collection status
Collected

plugins_6a7f93a24b3c81918caa0259f47b370a

Download plugin data (JSON)

Plugin details: freshly prepared. Prepared . Refreshed when plugin data changes or after 6 hours. Your account and Follow status are live.