Repository navigation
Description
NOT AN OFFICIAL MINECRAFT PRODUCT. NOT APPROVED BY OR ASSOCIATED WITH MOJANG OR MICROSOFT.
Planned from combsmith/worldchest Discussion #6 (graduated from combsmith/incubator Discussion #5). Source of truth: _bmad-output/planning-artifacts/prd.md (FR-1..FR-8) and brief.md §6–§7.
This epic is an evidence programme, not a build. It authorises no product code, Android/iOS app, backend, store listing or paid plan. The backup product (Share-sheet baseline + Shizuku Power mode + Desktop companion, sold as a ~$5/mo Family plan) appears only as the hypothesis under test. Almost every story is human-run (hands_off): interviews, hallway tests, device spikes and the Mojang question cannot be delegated to an agent.
Initiative success metric
The initiative has worked when each of brief.md §6 rows 1–5 carries a dated Result (a value, "not shown", or "not run" with a reason) and a decision comment checks them against Graduation criteria (a) and (b) and the "say no" conditions, within about five working days from the first test (SM-6), with zero lines of product code written (SM-C2). The evidence targets the decision reads against (PRD §8):
- SM-1: ≥3 of 5 interviewed parents describe a lost-world or device-change incident in the last 12 months.
- SM-2: fake-door conversion ≥ the pre-set Conversion threshold, counted only if visitors reach the pre-set floor (else inconclusive).
- SM-3: ≥2 of 3 non-technical parents reach Shizuku's "service running" screen computer-free and unaided.
- SM-4: a dated legal verdict exists, the Mojang question is sent, and the verdict is "permitted" ("prohibited" = no; "unclear" blocks approval).
- SM-5 (secondary): share-sheet round trip result recorded on both devices.
Counter-metrics (do not optimise): waitlist size bought with urgency copy (SM-C1); any product code (SM-C2); help given during a hallway test (SM-C3).
Cost cap
- Time: about one week (≈5 working days) of the owner's time in total (PRD §5, brief §7). FR-4 is time-boxed to one working day; FR-5 reading to about one hour.
- Money: $0 paid traffic, $0 counsel/legal fees (PRD §5, FR-1/FR-5 out-of-scope). Interview compensation, if any, is the Decider's choice and is recorded with the interview.
- Agent spend: none planned — all stories are hands-off human work; no dev-lead runs are expected.
Guardrails
- No child is a participant; no child data is collected. Interview notes name no child.
- No world files are uploaded anywhere.
combsmith/worldchestis public: commit only de-identified, aggregated results (counts, verdicts, device/OS versions). Never commit participant names, emails, phone numbers or raw interview notes.- Everything published follows the Naming rule and carries the Disclaimer (FR-6).
Dependency shape
Phase 1: brand-compliance checklist (FR-6) gates the fake-door launch (FR-1, whose thresholds are dated before the first visitor). Phase 2: FR-2, FR-3, FR-4, FR-5, FR-7 run in parallel. Phase 3: FR-8 results + decision, blocked by every test.
Untracked prerequisites
- Decider's dated "go" comment starting the evidence programme (PRD scope note) — the PRD reserves
idea:approvedfor the post-results decision, but incubator Discussion docs: seed planning artifacts from the incubator package (graduation) #5 already carriesidea:approved(2026-10-21); confirm the go is on record before the first test - Control of the
worldchest.appdomain plus a static host for the Fake door page (no evidence in the repo that it is registered or hosted) - An anonymous session-analytics provider for the Fake door (chosen and recorded before launch per FR-1)
- Access to one Android 15 device and one iOS device with Minecraft Bedrock installed and at least one test World (FR-4)
- A recruitment channel for five interview parents and three non-technical hallway-test parents with Android 11+ phones (PRD §9 Q2; owner's own network assumed)
- A private, non-repo location for raw evidence (interview notes, sign-up emails, hallway-test observations) — combsmith/worldchest is public
Planned from idea discussion #6 by the BMAD Scrum Master initiative-planner. Inert until a maintainer adds initiative:auto.
Activity
don-petry commentedon Oct 22, 2026
Owner decisions on the planner's open questions (2026-10-22)
idea:approvedon incubator Discussion docs: seed planning artifacts from the incubator package (graduation) #5 means "graduate and run the evidence programme". Applies to [Phase 3] FR-8: Results into brief §6 and the decision comment #15._bmad-output/evidence/, one file per story; no raw participant data in this public repo. Applies to [Phase 1] FR-6: Brand-compliance checklist for everything the programme publishes #8, [Phase 2] FR-5: EULA and Usage Guidelines note + written question to Mojang #9, [Phase 2] FR-3: Shizuku hallway test — three non-technical parents, unaided #12, [Phase 2] FR-7: Mojang native-backup watch (time-boxed, dated checks) #14.PRD updated in #16.