Idea to MVP evidence

Explore the idea before you ask it to become software.

Arcanum turns a rich prompt into a visible loop: refine the idea, author the definition and plan, route gaps, execute one bounded unit, and keep the evidence.

Start with a prompt See the Whisper example

Rich idea promptName the user, promise, constraints, unknowns, references, and ambitious future.

RefineExtract the core shape, compare routes, surface gaps, and choose the next owner.

Invoke define, design, planTurn intent into durable artifacts, design views, layering, and a work-pack.

Route blockersUse decision-gate, dispatch-spec, x-ray, or another refine pass when the plan reveals uncertainty.

Task-sessionExecute exactly one bounded task or SWU with scope, validation, and receipts.

Evidence and residueRecord what passed, what failed, what remains blocked, and where the next loop starts.

The exploration grammar

Whisper shows that every ambitious idea needs three cores.

In Whisper, the cores shape a text. In the broader Arcanum loop, they shape any tool, interface, research route, or workflow before implementation begins.

resonance_core

Idea promise

What human energy should the idea create? Relief, trust, ambition, clarity, wonder, confidence, or momentum.

relevance_core

Audience and domain fit

Who is it really for, why now, what world does it enter, and what objections must it respect?

trajectory_core

Transformation path

Where does the user start, what sequence carries them forward, and what evidence proves movement happened?

Worked specimen

Whisper is not only a writing tool. It is a pattern for exploring ideas.

The Whisper run began as an article idea about language and generative AI. It became a substrate, candidate tournament, composition plan, validator, drafts, task-session receipts, and publication residue.

Instead of saying "make this article better," Whisper asks what the piece should make the reader feel, why it belongs to its audience, and how the argument should move. Arcanum can ask the same questions of a frontend tool, a research workflow, a validator, a guide, or a new sigil.

Do not pick the first plausible route

Compare candidates across promise, fit, and movement.

The Whisper fixture preserves rejected alternatives. That matters because a rejected route can become a later transport, a different audience path, or a safer design after a blocker is resolved.

Executable language research note

Selected because it balances wonder, concrete Arcanum example, and reader agency.

Promise

Fit

Movement

opening pass citation pass audience pass

Anthropological bridge first

Strong public frame, but it risks opening with external authority instead of reader experience.

Promise

Fit

Movement

opening block citation pass audience pass

Practical prompting reframe

Useful and fast, but too likely to collapse into a prompting tutorial.

Promise

Fit

Movement

opening pass citation pass audience pass

From idea to parts

A selected route becomes responsibilities, dependencies, and checks.

This is where an idea stops being vapor. Each part gets a job, source inputs, must-do rules, must-not-do rules, and validation hooks.

Opening hookGround the reader in naming, language, or workflow before external authority appears.

Reference bridgeUse Harari as paraphrased context after the opening, with citation limits preserved.

Research contextFrame generative AI as a collaborator in symbolic system building.

Core insightShow language becoming executable through names, aliases, schemas, and tools.

Arcanum exampleTranslate sigils, spells, and schemas as public-facing examples.

ImplicationsConnect the pattern to builders, writers, researchers, and designers.

InvitationAsk the reader to name one workflow and begin shaping it.

When the loop finds a gap

Use the right capability for the shape of uncertainty.

refine

Use when the idea still needs shape, comparison, critique, repair, or a better next route.

invoke

Use when the idea needs durable definition, design, planning, layering, and work-pack artifacts.

decision-gate

Use when a consequential human choice blocks mutation or changes the direction of the system.

dispatch-spec

Use when research or execution needs multiple lanes, gates, handoffs, and evidence ownership.

x-ray

Use when the hidden structure, dependencies, lifecycle, or relationships are hard to see.

task-session

Use when one task or SWU is ready enough to execute with validation receipts.

Use it on your own idea

Start with the messy version. Arcanum can shape it.

The first prompt is a seed, not a final brief. Give the system enough material to protect the ambition while finding the first executable proof.

[$refine] I want to develop this idea:

Describe the users, promise, current workflow, what feels missing, references, constraints, risks, unresolved decisions, MVP proof, and the ambitious version.

Please refine this into definition, design, plan, gaps, blockers, and the next executable task.

[$dispatch-spec] Create a research route for this idea.

Streams: source truth, market practice, implementation, risks, validation or fixtures.

Output: source ledger, claim map, implications, blockers, and next work-pack.

[$task-session] Execute <work-pack path> --swu <SWU-ID>

Use only when scope, source links, write boundaries, blockers, validation, and completion evidence are clear.

Evidence ladder

The closer you get to mutation, the heavier the evidence should become.

Prompt evidence

User, pain, promise, constraints, unknowns, examples, and success signal.

Design evidence

Definitions, views, trade-offs, candidate comparisons, and rejected alternatives.

Execution evidence

Work-pack, SWU, write scope, validation commands, receipts, and changed files.

Residue evidence

Blocked claims, next routes, human review needs, and reusable lessons.