nomue (earlier version) · Archived

Archived nomue roadmap

Archived nomue roadmap, last updated September 19, 2026. nomue was Licklider’s earlier product, and its verification capabilities are now part of SciGround. This page records what nomue had released and what it planned at that date; it is not a current plan.

At the time, the roadmap separated public artifacts, release candidates, limited releases, public discussions, and adopted plans. Its stages did not promise ship dates.

What happened next

nomue’s verification work continues in SciGround

No decision has been published on the Release 2–5 or Welch successor proposals, and the public nomue Protocol repository records no change after September 28, 2026. Release 1 is unchanged. nomue’s verification work now continues in SciGround.

Several method families that nomue planned are now supported analyses in SciGround, including paired before-and-after comparison, rank comparison of two independent groups, and comparison of three or more groups.

nomue is now part of SciGround · Analyses SciGround supports · Connect SciGround

Released by September 2026

A public trust layer for Welch verification

Anyone could install the public @licklider/nomue-verifier package from npm and run nomue verify locally to check a conforming Release 1 Record for independent two-group continuous outcomes under the two-sided Welch two-sample t procedure. It recomputed the covered numerical quantities and returned a machine-readable report of the scoped checks without calling a nomue server after installation.

nomue Protocol Release 1 Public Draft

Public Record, bundle, check, and result meaning for the first supported scope.

Protocol source

@licklider/nomue-verifier

Open-source local verification, published on npm as 0.2.1-rc.1 and exercised across three operating systems and two Node.js versions.

npm package

Released — release candidate

Agents could call nomue Record Verifier over MCP

@licklider/nomue-verifier-mcp@0.2.0-rc.1 was published on npm and in the official MCP Registry as nomue Record Verifier. It exposed the method-neutral verify_nomue_record tool over local stdio, limited to Release 1 Welch Record verification, and delegated to @licklider/nomue-verifier@0.2.1-rc.1. It required no account, API key, environment variable, or Licklider-hosted service, and it did not add paired-t, Wilcoxon, Mann–Whitney, method selection, raw-sample calculation, or an overall scientific verdict.

0.2.0-rc.1 stdio verify_nomue_record 3 OS × 2 Node versions

MCP guide (earlier version) · npm · Official Registry · CI runs

Limited release and public proposals

Managed Welch verification and proposed next capabilities

No decision has been published on the Release 2–5 or Welch successor proposals, and the public nomue Protocol repository records no change after September 28, 2026. Release 1 is unchanged. nomue’s verification work now continues in SciGround.

Managed agent-facing Welch workflow — limited Release 1

Approved recipients could submit data and required scientific declarations for a supported Welch calculation, or a claimed result with structured evidence for checking, through authenticated MCP and HTTP. The service returned scoped outcomes, reasons, evidence, and next actions. Public self-registration was not offered, and hosted results were not emitted as public Records for local replay.

Release announcement

Paired-t / Protocol Release 2

The paired-t candidate reached an independently reviewed decision-readiness packet with its numerical and execution evidence, review results, and Release 1 safeguards. No decision on it was published.

SciGround now covers paired before-and-after comparison with a paired t-test.

Public RFC · Before and after in SciGround

Independent groups and multiple comparisons / Protocol Release 3

The proposal made design, comparison families, result meaning, and error-control questions explicit across 49 catalogued procedures. Its evidence scope was limited to supplied originals; method adoption and numerical support were separate decisions.

The public discussion began on September 9, 2026; the earliest decision date was October 9, 2026. No decision was published.

Scope and evidence · Discussion record

Two-factor experiments / Protocol Release 4

The proposal covered balanced two-by-two fixed-factor experiments. Its unissued numerical, report and controlled-execution candidate reached independently reviewed final readiness, without establishing Protocol support. A separate amendment covered strict binary64 comparison and completed indeterminate results.

The public discussion began on September 9, 2026, with an earliest decision date of October 9, 2026; the amendment, opened on September 18, 2026, had an earliest decision date of October 18, 2026. The formal decision packet completed its review. No decision was published.

Scope and evidence · Discussion record · Numerical scaling study

Shared design and timing evidence / Protocol Release 5

The proposal covered a common evidence view for declared study design and selection timing across analysis families: versioned mappings, timing declarations, explicit limits on what a passing check means, and a shared report view. Its initial scope considered independent two-group, paired two-condition and independent multi-group continuous outcomes, each depending on its own separately accepted successor.

The public discussion began on September 17, 2026; the earliest decision date was October 17, 2026. No decision was published.

Scope and evidence · Discussion record

Controlled agent evaluation

The evaluation program examined nomue through decision quality, cost and time, and concrete cases, with explicit comparisons and linked reproduction materials. It published two developer-led preprints on constructed Welch workflows and selected cases from the whole-submission study.

nomue evaluation (earlier version)

Planned platform evolution (archived plan)

Planned: make each verification call portable, persistent, and reusable

In September 2026, nomue had started work on assembling and emitting Records, and its adopted plan targeted an invite-free release covering Welch, independent multi-group and paired two-condition capabilities, with no promised date. No outcome of this plan was published under the nomue name.

01

Independently verifiable Records

Carry exact capability, engine, Protocol, and result evidence into a Record that another party can check without trusting the product runtime.

02

Agent-native Project state

Let an authorized agent return to the same research task without rebuilding canonical context from chat history.

03

Persistent resumability

Reconnect and continue long-running research work while keeping project truth separate from conversational memory.

04

Generalized capability kernel

Stabilize the common contract, result, refusal, evidence, and version surfaces after materially different methods prove what is shared.

Planned scientific expansion (archived plan)

Ten method families were planned

The nomue plan extended beyond independent two-group Welch into the following research structures. Each family was to earn its own scope, evidence, and release. Where SciGround now covers a family, the entry says so.

Multi-group preparation included reviewed, accepted source characterizations of unequal-variance intervals, control and best-treatment comparisons, and closed-testing graphs. These results clarified the assumptions and outputs a future rule would have to preserve; they did not select or release those methods. Unequal-variance comparisons · Control or best? · Testing graphs.

Longer-term plan (archived)

Planned: real boundary cases become governed improvements

nomue planned a governed learning loop that would turn privacy-minimized failures and boundary cases into reviewed evidence, regression cases, and improved future capability versions, without treating model self-judgment as scientific ground truth.