Development roadmap

MNT01 · DEPENDENCY-VERIFIED PRIORITIES

Advance only what its dependencies can support.

This queue separates executable work from work that needs a particular environment, operating capability, or independent reviewer. Rank never overrides a safety gate.

Selection rule

Protect active boundaries, prefer satisfied dependencies, and preserve a recoverable release after every increment.

Stop rule

Missing reviewers, operating capacity, or test environments are dependencies—not software defects to bypass.

01BLOCKED · REVIEWERS

EXP04

Verified researcher, reviewer, and institution roles

Least-privilege investigator, laboratory, data-steward, method, statistics, safety, ethics, replication, adjudication, and maintenance roles with conflicts, separation of duties, expiry, and appeal.

Depends on
EXP03A · P18H
Blocked by
verified identity provider · institutional verification method · named operating owners · completed authentication recovery and revocation exercises · independent privacy legal accessibility and appeal review · completed P18H evidence
Stage hooks
identity.provider.changed · experiment.role.changed · privacy.expectation.changed
Propose evidence
02BLOCKED · ENVIRONMENT

UX06B2B2

200% text-enlargement verification

Runtime evidence across the five public task journeys, with one evidenced repair at a time.

Depends on
UX06B2B1
Blocked by
browser environment capable of verified 200% text enlargement; the 2026-09-16 managed cloud browser gate check exposed no effective browser-zoom control
Stage hooks
accessibility.expectation.changed · usability.interaction.changed
Propose evidence
03BLOCKED · OPERATING CAPACITY

EDU01S2B

Testable filtering and quarantine implementation

Tested production personal-information filtering, unknown-link and file quarantine, reporting, expiry, appeal, and failure-safe shutdown while student interaction remains disabled.

Depends on
EDU01S2A
Blocked by
approved interaction architecture · named safeguarding owner · operating capacity
Stage hooks
student.interaction.changed · minor.safeguard.changed · platform.access.changed
Propose evidence
04BLOCKED · REVIEWERS

EDU02

Independent educator and safeguarding review

Dated review evidence and a supervised-pilot decision; no automatic activation.

Depends on
EDU01S2A · EDU01S2B
Blocked by
named educators · independent safeguarding reviewer · accessibility reviewer · applicable legal review
Stage hooks
student.workspace.changed · minor.safeguard.changed · accessibility.expectation.changed
Propose evidence
05

P18H

Independent forum expansion decision

Explicit expansion, restriction, or rollback decision with limitations, staffing, failure evidence, and independent review.

Depends on
P18G · P18I
Blocked by
verified expanded operating capacity · consent-aware outbound email provider · privacy-preserving redirect-chain resolver and replaceable link reputation provider · independent privacy, security, legal, accessibility, backup, archival, and appeal review
Stage hooks
forum.policy.changed · notification.delivery.changed · link.safety.provider.changed · moderation.capacity.changed · appeal.process.changed
Propose evidence
06BLOCKED · DEPENDENCY

P19

Shared project and grant-management workspace

Staged collaboration workspace separating public summaries from restricted proposals and financial records.

Depends on
P18H · P11 · P12 · P14
Blocked by
P18 controls · financial and funder compliance · least-privilege project access · audit and handoff controls
Stage hooks
platform.access.changed · storage.schema.changed · retention.schedule.changed
Propose evidence
07BLOCKED · DEPENDENCY

P21B

Opt-in AI organization and diverse inspiration pool

Reversible per-item AI organization and diversity-aware inspiration retrieval that never treats popularity, identity, or reuse as scientific authority.

Depends on
P21A · P18H · EDU02
Blocked by
authenticated owner-scoped storage · per-item consent and withdrawal · personal-information and secret filtering · moderation and abuse response · retention deletion and licensing review · minor exclusion or independently reviewed safeguarding · independent privacy legal accessibility and appeal review
Stage hooks
platform.access.changed · storage.schema.changed · privacy.expectation.changed · retention.schedule.changed · forum.policy.changed · student.interaction.changed
Propose evidence

MAINTENANCE PAUSED · REVIEW MNT12C

Every remaining maintenance item now requires external review, a real change trigger, or nonproduction operating capacity. Maintenance cannot satisfy a reviewer, safeguarding, access, or operating-capacity gate.

01BLOCKED · REVIEWERS

MNT12C

Authority, funding, and safeguarding review

Record dated, expiring decisions for authority, privacy, age suitability, eligibility guidance, and funding currency while preserving quarantine and non-approval boundaries.

Depends on
MNT12A · MNT12B · MNT12CP · MNT12CV
Blocked by
named source-authority and privacy reviewers · named safeguarding reviewer for student destinations · named funding-currency reviewer for grant destinations
Stage hooks
funding.opportunity.changed · student.external_link.changed · privacy.expectation.changed · minor.safeguard.changed
Propose evidence
02BLOCKED · DEPENDENCY

MNT13

Dependency-upgrade rehearsal

Rehearse one triggered dependency change with compatibility, build, interaction, compute, rollback, and replacement evidence before production adoption.

Depends on
MNT02 · MNT10 · MNT11
Blocked by
an actual support-window, security, or compatibility trigger · isolated upgrade branch and rollback evidence
Stage hooks
dependency.major.changed · platform.hosting.changed · limit.compute.changed
Propose evidence
03BLOCKED · OPERATING CAPACITY

MNT14

Access and hosting recovery exercise

Verify audience restoration, least privilege, credential recovery, deployment rollback, and incident handoff without exposing private data or changing production access.

Depends on
MNT02 · MNT06 · MNT10 · MNT11
Blocked by
separate authorized owner and reader test roles · provider recovery capability · privacy-safe nonproduction exercise environment
Stage hooks
platform.access.changed · platform.hosting.changed · privacy.expectation.changed
Propose evidence

RECENTLY COMPLETED

P18I · Profile, following, notification-preference, and link-safety foundation

Account-scoped profiles now consolidate explicit topic, subtopic, and post follows; separate email selections; child mute exceptions; authored activity; and profile editing without exposing account email. Profiles, threads, and replies share one fail-closed link intake. Submitted destinations remain inert until staged redirect-chain and reputation checks pass; moderators may disable links but cannot activate or bypass them. Outbound email delivery is not enabled.

MNT01 · Development-state consistency layer

P13, P14, and P15 continuation labels corrected to implemented; future priority claims checked against this queue and the maintenance registry.

EDU01S2A · Synthetic minor-interaction safety fixtures

Eleven synthetic cases now test reference handling for personal information, precise location and schedules, approved and unknown links, shortened URLs, files, off-site contact, safety reports, and ambiguous numeric patterns. No interaction channel was enabled.

UX06B2B2P · Portable 200% enlargement verification protocol

A versioned evidence contract now defines the measurable entry gate, five public journeys, result fields, privacy boundary, hooks, and re-entry path. The test itself remains blocked and unclaimed.

UX06B2B2V · Fail-closed enlargement evidence validator

Six synthetic cases now verify the portable evidence format, reject missing measurements and journeys, reject CSS substitutes and privacy violations, and prevent unmeasured pass claims. UX06B2B2 remains blocked.

MNT02 · Platform compatibility and obsolescence register

A versioned registry and automated checker now detect drift in runtime, package manager, framework, storage bindings, access assumptions, change hooks, and replacement requirements without performing automatic upgrades.

MNT03 · Feature interconnection registry

Eight major capabilities now map to their routes, source surfaces, public artifacts, feature dependencies, maintenance hooks, checks, limitations, and successor paths; missing or stale connections fail the standard site check.

MNT04 · Maintenance-hook dependency validation

Thirty-nine evolution hooks now undergo unique-identifier, known-prerequisite, cycle, required-response, stage-gate, stop-condition, and successor-path validation while blocked product stages remain closed.

MNT05 · Public artifact lifecycle registry

Every public JSON artifact now belongs to one dependency domain with a lifecycle class, maintenance hooks, public nonpersonal boundary, retention rule, same-release change rule, and successor path; additions, removals, renames, and orphaned files fail the standard site check.

MNT06 · Route lifecycle and access-boundary registry

All public pages and API routes now belong to dependency domains with explicit access, interaction boundaries, maintenance hooks, and successor paths; route additions, removals, renames, source mismatches, or orphaned handlers fail the standard site check.

MNT07 · Dependency-ordered verification control registry

Every release checker now declares its dependencies, protected scope, evidence class, failure policy, hooks, limitations, and successor path; missing, unwired, reordered, or silently disabled checks fail the standard site check.

MNT08 · Change-impact routing contract

Every tracked source file now routes through at least one rule that identifies required hooks, controls, update surfaces, stage gates, limitations, and stop conditions; unrouted files or stale impact fixtures fail the standard site check.

MNT09 · External dependency and link lifecycle registry

Every literal HTTPS destination now belongs to a use class and exposure state with authority, audience, privacy, review, and replacement boundaries; additions, removals, duplicates, unsafe active fixtures, and unregistered references fail the standard site check without claiming live network availability.

MNT10 · Maintenance cadence and evidence freshness control

Per-release, monthly, quarterly, semiannual, and annual maintenance stages now declare controls, hooks, owner classes, evidence types, authority limits, review dates, missed-run responses, and successor paths. A separate maintenance queue advances MNT11 without reopening externally blocked product stages.

MNT11 · Clean-room recoverability drill

A new nonproduction checkout recovered the registered remote source, restored pinned dependencies through the project-owned install path, passed all controls and the production build, and produced a validated but undeployed hosting package. The evidence records package integrity, recovery findings, and explicit limits on access, data, provider, and production recovery claims.

MNT12A · Portable external-review evidence protocol

A public protocol and ten synthetic cases now separate source inventory, network observation, and named authority, funding, and safeguarding review. The fail-closed validator rejects privacy violations, nonpublic references, unrecognized cross-origin redirects, stale observations, and missing reviewer roles without approving any destination.

MNT12B · External-destination network observations

Thirty active destinations now have dated, nonpersonal observation records, declared transport limits, expiry, and named-review handoffs. The protocol now permits only DOI-preserving citation-resolver redirects while retaining non-approval and quarantine boundaries; MNT12C remains blocked on named reviewers.

MNT12CP · Named-review worklist preparation

Five priority-ordered batches now cover all thirty external observations with required reviewer roles, checks, finding questions, per-URL decision fields, expiry, privacy limits, and successor rules. Every review and decision field remains empty, so MNT12C stays blocked until named reviewers act.

MNT12CV · Fail-closed human-review evidence validator

Twelve synthetic cases now accept only complete, per-destination, role-complete, nonpersonal, unexpired records and reject missing authority, funding or safeguarding roles, personal information, stale or absent expiry, group-level approvals, unknown destinations, invalid decisions, and incomplete records. No live review was performed.

P20A · Consent-gated contributor registry with grant matching

A public, initially empty categorized registry now defines evidence, review, significance, privacy, consent, attribution-order, correction, withdrawal, and CV-citation requirements. At recognition it also requires a dated official-source search for five related current grant opportunities, eligibility routes, match rationales, pre-application rechecking, and a documented shortfall rather than padded results. Eleven synthetic cases enforce the combined boundary; no real person or grant search was fabricated.

P18F2 · Groupthink-resistant effort introductions

A read-only forum surface, portable contract, and thirteen synthetic cases now require immutable independent positions, work-based matching, meaningful divergence, mutual opt-in, comparison charters, preserved branches, perspective rotation, neglected-area protection, recognition and funding separation, and teacher- or guardian-mediated minor boundaries. Matching, identity disclosure, messaging, and contact exchange remain disabled.

P21A · Device-local moment-of-inspiration capture and portable AI workspace

A persistent adult-workspace dock recovers an unfinished draft, sends a thought to a private browser ToDo list, or prepares one portable interaction workspace. The explicit ChatGPT handoff copies the prompt and opens the official origin without placing content in a URL; users review, paste, and send it. The local workspace supports optional response paste-back, participant labels, same-origin forum topic/post links, JSON export/import, and device sharing. Labels grant no access, and automatic submission, response retrieval, server storage, live invitations, private messaging, real-time editing, student routes, and the cross-user inspiration pool remain disabled.

P22A · Role-specific human and AI bibliographic attribution

Bibliographic annotations and contributor examples now preserve original source authorship while separately crediting James Dunn for originating concepts and program direction and OpenAI ChatGPT for documented synthesis, drafting, cross-reference organization, software implementation, or editorial coordination. The contract rejects claims of independent verification, scientific authority, or source authorship from AI participation.

EXP01 · Prioritized experimentation development register

A public machine-readable roadmap now orders twelve experimental-research improvements from claim-packet preparation through identity, preregistration, safety, laboratories, raw evidence, analysis, replication, claim promotion, utility, federation, and recovery. Fourteen evolution hooks preserve dependencies and stop automatic progression before external authority or evidence is required.

EXP02 · Portable experimental claim packet

A public non-collecting contract now preserves identity, question, candidate relation, conventional and null comparisons, detector and frame, preregistration, safety, runs, raw evidence, analysis, review, replication, attribution, AI assistance, claim decisions, corrections, withdrawals, licensing, and exchange. It cannot authorize physical execution or scientific promotion.

EXP03 · Fail-closed claim-packet readiness fixtures

Fifteen synthetic cases now accept complete E0, E2, E3, and externally authorized synthetic E4 structures while rejecting missing comparisons, missing predictions, AI self-approval, popularity promotion, Site-generated execution authority, mutable raw evidence, private contact fields, unsupervised minors, unsupported cross-site maturity, and missing export provenance. No live experiment or approval was represented.