Research journey · frame + predict · browser-local

Start with one question and make the test falsifiable.

Choose a strong protocol scaffold or begin blank. Draft the question, held-fixed boundary, disconfirming observation, qualification rule, and smallest test; then carry them as deterministic JSON. Nothing is uploaded or executed, and the bundle cannot lock a protocol, admit evidence, or grant publication approval.

Commit phaseThree-step local draftDeterministic JSON v2No implicit approval

Protocol starters · evidence none

Choose a study scaffold, then make every field your own.

These examples preserve strong research patterns from recent protocol work. They do not activate a study, call a model, or turn a draft into evidence.

Build from an empty question

Blank falsifiable study

Name one comparison, one held-fixed boundary, one falsifier, and the smallest test that could change a decision.

Protocol starter · Evidence none · No run performed

Use this starter

Behavioral alignment

Behavior change, not wording

Compare action-grounded behavior against matched language-only and sham controls while protecting collateral utility.

Protocol starter · Evidence none · No run performed

Use this starter

Multi-principal safety

Permissioned handoff integrity

Test bilateral permissions and content-addressed lineage against metadata-only logging, then measure bad-artifact acceptance and useful completion.

Protocol starter · Evidence none · No run performed

Use this starter

Reproducibility

Reproduce a claimed effect

Specify an exact independent reproduction against the released environment, primary output, tolerance, and decision rule.

Protocol starter · Evidence none · No run performed

Use this starter

One study draft · three working steps

Frame, specify the test, then review and export.

The workbench format-checks a detached local draft. It does not judge the science, establish independent review, or create research progress.

01Frame the claimQuestion, intended comparison, and current claim boundary.
02Specify the test before evidenceFalsifier, held-fixed conditions, decision rule, and smallest protocol.
03Review and exportBlocked claims, unknowns, next human gate, readable draft, and deterministic JSON.

Protocol: Draft
Evidence: None
Review: None attached
Execution: External · not authorized
Publication: Not authorized

Browser-local boundary · no run, evidence, review, approval, or publication

Known protected Atlas markers and secret-like values are rejected, but the heuristic scan cannot classify every sentence. Do not paste private material. Imports and exports contain declarations only; artifact bytes are never embedded, and a format check grants no authority.

Step 1 · Frame the claimTitleQuestionHeld-fixed boundary

Step 2 · Specify the test before evidenceExperiment titleComparison / interventionDisconfirming observationDecision ruleSmallest test

Step 3 · Review and export

What may be said now: a detached protocol draft exists. Evidence, independent review, execution, approval, and publication remain absent.

Intended study summaryLatest draft update

Intended claims, not supported claimsBlocked claims, one per line

Publication boundaryNext human gate

Technical detailsPortable identity, object links, declared artifact metadata, and unsigned annotations

Bundle identity and object referencesPortable bundle id

Selected objects

Declared artifact references

One per line: name | media type | byte size | optional lowercase SHA-256 | note. These are unsigned declarations, not admitted evidence. File bytes are not embedded.

Declared artifact metadata

Unsigned annotations

One per line: claimed attribution | open or addressed | note. A typed name is unverified and never establishes independent review.

Portable unsigned annotations

No study is loaded. A new draft exists only on this page and is lost on navigation or refresh unless you export it.

Export blocked

$.ideaBrief.title Expected non-empty text no longer than 240 characters.
$.ideaBrief.question Expected non-empty text no longer than 8000 characters.
$.ideaBrief.currentClaimBoundary Expected non-empty text no longer than 8000 characters.
$.experimentCard.title Expected non-empty text no longer than 240 characters.
$.experimentCard.hypothesis Expected non-empty text no longer than 8000 characters.
$.experimentCard.falsifier Expected non-empty text no longer than 8000 characters.
$.experimentCard.decisionRule Expected non-empty text no longer than 8000 characters.
$.experimentCard.protocol Expected non-empty text no longer than 8000 characters.

Readable handoff

Local study draft

Protocol draft · Evidence none · No review attached · No run performed

Untitled study draft

Question: Not yet stated.
Current claim boundary: Not yet stated.
Comparison or intervention: Not yet stated.
Disconfirming observation: Not yet stated.
Decision rule: Not yet stated.
Smallest test: Not yet stated.
Next human gate: Complete the required commitments, then obtain independent methods review before any protocol lock or execution.

Current permissible statement

A detached protocol draft exists. No empirical evidence, protocol lock, independent review, execution, approval, or publication authority is attached.

Technical representation · inspect schema-normalized local JSON