MIRR TEST - Checklist

Checklist unavailable

This page needs verified TEST configuration for this origin. No test controls are enabled.

Current TEST candidate

Batch
Source commit
Origin
Processing and data

Stub output cannot establish real-model quality. This page makes no API, account or provider requests.

How a batch reaches release

Record this batch and source, test it, and review the results. Jason explicitly greenlights the same batch/source before a separate production release. Changed source needs review again. No checklist result grants approval or deploys anything.

Your test context

Keep Android and desktop runs separate. Use an ordinary URL; no USB connection is needed.

Device Operating system Browser How MIRR is open

Data mode Synthetic examples Owner-selected snapshot - transfer flow not implemented

Fictional captures

These are fictional situations, not diagnoses or labels for real people. Edit the raw notes or add your own synthetic holdouts. Do not paste private journals.

Scenario Reflection tone comparison - record only

This selector records a comparison choice only. To change your actual saved Daily style, open MIRR Settings, then Edit saved Daily style. Gentle, Direct and Blunt but respectful are supported; Plain and Source are reference choices. Weekly uses a fixed style. Stub output cannot test model quality.

Two human-written holdouts

Use your own fictional wording. Human review is still required; generated cases do not grade themselves.

Reviewer notes - excluded from copied texts

These observations apply only to the unchanged example. Edits and holdouts need fresh human judgment. Labels and reviewer notes must stay out of model input.

Look for

Claims the evidence does not support

Text-only JSON for manual copying

Only nonempty capture and holdout texts are copied. Scenario names, dates, reviewer notes and tone are excluded. Copying does not create captures or run MIRR.

Dated Weekly fixture plan

Prepare seven completed MIRR days for a future approved TEST import. The plan keeps fictional capture text and local-noon timestamps; it excludes holdouts, scenario labels, tone and reviewer guidance. Downloading imports nothing and approves no capture.

Last completed MIRR day UTC offset minutes, west positive

Browser checks

Not run: untested. Pass: expected behavior observed. Fail: behavior differed. Blocked: a prerequisite is unavailable. Use only the disposable TEST accounts/actions approved for this batch.

Physical-device checks

Browser emulation cannot prove installation, phone delivery or lock-screen capture. Keep each device/browser result separate.

Keep your results

0 of 14 checks recorded. Results are manual observations.

Results stay in this tab and disappear when it closes or reloads. There is no autosave or upload. Download explicitly to keep a report; restore only a report for this exact origin, batch and source.

Use short expected-versus-actual notes without private journals or account details. Reports exclude fixture and holdout text. Restore does not restore those edits.

Restore results JSON

Maximum report size: 32 KiB. Invalid or mismatched reports leave current results unchanged.