Release roadmap & verification

Staged open-source releases with explicit evidence grades.

Project Meno is preparing its reference components for public open-source release in waves ⁠— publishing what has been tested, what remains experimental, and what awaits human and external security review.

Staged waves

Curated, clean-history public repositories.

Rather than dumping an uncurated laboratory tree, Project Meno’s release preparation separates self-contained open-source components into clean repositories with verified third-party font and dependency attributions, secret scanning, and reproducible test suites:

  • Wave 1 (in preparation): Machine-native principles specification, core .mnd and .mnv parser and slicer tools, finite-state .mni policy compiler and exhaustive checker, and context-slicing agent workflow utilities.
  • Wave 2 (planned): MNFS content-addressed object store, content-blind relational whiteboard index, and reproducible filesystem benchmark harnesses.
  • Wave 3 (longer-horizon research): Experimental wire-reconciliation sketches, native compilation experiments, and active-inference micro-node research tracks.

Evidence ledger

Distinguishing measured systems from prototypes and paper models.

Every public benchmark or architectural claim in a Project Meno release is required to carry one of three explicit evidence grades so readers never have to guess maturity:

Grade A · Measured
Tested against real operating-system primitives, external repositories, or established baseline tools with reproducible scripts and raw receipts.
Grade B · Prototype
Working reference code validated on synthetic corpora, local loopback interfaces, or controlled test harnesses.
Grade C · Analytical
Formal or mathematical design notes modeled on paper and not yet implemented or empirically benchmarked.

Code & security review

AI authorship disclosure, human review, and external audits.

Transparency about how code was produced matters as much as benchmark numbers:

  • AI-assisted implementation disclosure: Reference code in the Meno research lab has been generated and cross-reviewed using frontier AI models under James Everson’s design specifications and spot checks, and has not yet undergone complete human line-by-line inspection.
  • Component review ledger: Each published module will record its current review state ⁠— AI cross-review, maintainer logic spot check, systematic human line-by-line review, or external audit ⁠— so downstream users know exactly what has been inspected.
  • Planned independent security review: Before any Meno parser, cryptographic boundary, or policy gate is recommended for high-trust production environments, JREonline LLC plans an independent third-party secure code review and penetration test scoped to risk-bearing components.

Next: Adoption, consulting & contact →