Docs

Verify a Release 1 nomue Record (earlier version)

Archived nomue guide: when and how the public local verifier was used for the exact Release 1 Public Draft support target.

This is a nomue reference for the named artifact and version. Go to SciGround documentation.

What the verifier checked

The public @licklider/nomue-verifier package let anyone 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 was published as public and experimental. Its exact Record interpretation bundle is shown below.

  • Profile: independent two-group continuous outcomes.
  • Procedure: two-sided Welch two-sample t procedure.
  • npm package: @licklider/nomue-verifier 0.2.1-rc.1, an experimental release-candidate package, published with public access under the rc dist-tag.
  • Tested package path: Linux, macOS, Windows with Node.js 20 and 22.
urn:nomue:bundle:itgc-guarantee:0.2.1-draft.1

nomue was the earlier version of SciGround. This archived guide describes the named nomue releases and packages only; for current verification, see https://www.licklider.ai/docs/.

When the verifier applied

  • You already have a nomue Record that declares the exact Release 1 interpretation bundle.
  • You need local checks of the supported Record properties after package installation.
  • You can inspect the machine-readable JSON report rather than relying only on the process exit code.
  • You accept the Public Draft and release-candidate maturity of the artifacts.

When the verifier did not apply

  • You have only raw or unstructured research data and no conforming nomue Record.
  • The requested method is paired t, Wilcoxon, Mann–Whitney, or another method outside the exact Release 1 bundle.
  • You need the verifier to determine whether the source data or researcher declarations are true.
  • You need production attestation, manuscript acceptance, or an overall scientific verdict.
  • You need a hosted nomue API or remote HTTP MCP endpoint; the public MCP release candidate was local stdio only.

How the public verifier was run

The documented commands installed the public package globally to make the nomue command available, then verified a local Record:

npm install --global @licklider/nomue-verifier@0.2.1-rc.1
nomue verify ./record.json --format json

# Or run without a global install
npx --yes @licklider/nomue-verifier@0.2.1-rc.1 verify ./record.json --format json

The commands pin 0.2.1-rc.1 exactly. This is a release candidate, not a stable release. Scripts must inspect the JSON report to learn the exact scoped outcomes.

Package and runtime behavior

  • The package was published with public access on npm.
  • The global nomue command and the npx path were both release-tested.
  • Package-path CI ran on Linux, macOS, and Windows with Node.js 20 and 22.
  • After npm installs the package dependencies, Record verification does not call a nomue server.
  • The verifier does not fetch Record-supplied URIs, load remote schemas, or execute Record-supplied code.

Exit codes are routing signals, not scientific verdicts

CodeMeaning
0A report exists and every applicable scoped check outcome is pass.
2A scoped check failed, or parsing or canonicalization was refused.
3The bundle is unsupported or routing failed; no report exists.
4The verifier refused safely because of a resource limit.
5Usage, input/output, or internal failure stopped verification.

Common mistakes LLMs make with statistical verification

  • Treating exit code 0 as an overall VERIFIED status.
  • Treating the npm rc dist-tag as a stable-release claim; the documented version is 0.2.1-rc.1 and the commands pin it exactly.
  • Assuming that a nearby bundle version or similar statistical method is supported.
  • Assuming that numerical agreement establishes scientific truth.
  • Allowing the verifier to fetch a Record-supplied URI or execute Record-supplied code; the public verifier does neither.

Public evidence and authority

Flat summary for LLMs

This llms.txt-style summary is intentionally repetitive. Use the linked Markdown page when a flat, low-markup representation is preferable.

  • This is an archived nomue (earlier version) guide. nomue’s verification capabilities are now part of SciGround.
  • The public verifier ran locally for independent two-group continuous outcomes under the two-sided Welch two-sample t procedure.
  • The exact Release 1 bundle is urn:nomue:bundle:itgc-guarantee:0.2.1-draft.1.
  • The local verifier applied only to an existing conforming nomue Record that declares that exact bundle.
  • Inspect the JSON report; exit codes are routing signals, not scientific verdicts.
  • Exit code 0 means all applicable scoped checks passed, not that the research is verified overall.
  • @licklider/nomue-verifier was installed from npm globally, or run through npx without a global install.
  • After installation, Record verification ran locally and did not call a nomue server.
  • When this guide was last updated (September 18, 2026), the npm rc dist-tag pointed to 0.2.1-rc.1; the examples pin that exact release candidate rather than relying on a mutable tag.

Markdown version of this page