AA b14: Adversarial review of the Zenodo mirror page#

VVN: aa-adversarial-zenodo-dv_ClaOp47Max_NNv1_2026m04d25
Mode: EDEN
Effort: Max
Status: DONE (executed 2026m04d26 by Claude Opus 4.7 Max in EDEN/max session paired with AA b13; findings drove the SHA256 manifest, source-tree commit recovery, license harmonization to CC BY 4.0, SDM citation chain correction, toolchain version pinning, and the Zenodo description revision deposited 2026m04d27)
Origin: Created 2026m04d25 during the Evolvix bridge session. The Zenodo mirror page (sta1-evx/evx-prototype-compiler.rst) is mostly LLoL’s authoritative Zenodo wording; Claude restructured it for RST and added a Mirror notice + downloads section. An adversarial read on this page is genuinely reviewing LLoL’s scientific writing on Evolvix — a different kind of work than the bridge review and deserving its own session.

What this AA proposes#

A fresh Claude session reads the Zenodo mirror page at Prototype Evolvix Compiler — Zenodo Archive of Past Foundations adversarially from two angles:

  1. Hostile reviewer for scientific reproducibility — a journal/repository reviewer assessing whether the deposit and its description meet the standard for a citable scientific software archive. Are the claims about mathematical reliability supported? Is the version string trustworthy? Are the platforms documented sufficiently for reproducibility? What would a strict reviewer demand before accepting the deposit?

  2. Hostile language-design expert — a senior PL researcher who reads the Implementation, Mathematical Reliability, and “Prototype” qualifier sections critically. Are the design decisions defensible? Is the “1:99 best-idea-rejection ratio” supportable? Does the “Prototype” qualifier honestly describe the limitations? Where would an expert spot over-claiming or under-disclosing?

For each angle, surface concrete holes and propose specific fixes (with line numbers and replacement text), so LLoL can apply or reject each suggestion individually.

Why a separate session?#

The original session that produced the Zenodo mirror page (Claude Opus 4.7 Max, 2026m04d24) is biased: that Claude restructured the content, decided where to put the Mirror notice, drafted the Downloads section, and helped shape the framing of the citation block. A genuinely critical adversarial read needs a Claude that has not made those choices.

Even more importantly: most of the substantive content is LLoL’s authoritative Zenodo wording. An adversarial review on it is reviewing LLoL’s scientific writing, not Claude’s restructuring. That deserves a fresh, focused session where the reviewer is not entangled with “but Claude already restructured this.”

The Zenodo deposit itself (DOI 10.5281/zenodo.19679456) should also be checked — the mirror is supposed to faithfully reproduce the deposit; any drift between the two is a real bug.

Execution prompt (paste into a fresh Claude session)#

Dedicated session: Adversarial review of the Prototype Evolvix Zenodo mirror page and the actual Zenodo deposit text. Mode: EDEN. Effort: max.

The page is at: source/good-news-pack/vv/mmv3/flyingscroll/transwarpkey/sta1-evx/evx-prototype-compiler.rst

The Zenodo deposit is at: https://doi.org/10.5281/zenodo.19679456

This page is mostly LLoL’s authoritative Zenodo wording, restructured into proper RST by Claude on 2026m04d24. Three things must be reviewed: (a) the substantive scientific content, (b) the structural choices the restructuring made, (c) any drift between the local mirror and the actual Zenodo deposit (compare wording).

Pre-read: - source/_POST/DD/b/11/b11-trifecta-italics-no-commas.rst - source/_POST/DD/b/12/b12-ladder-pedagogy-for-acronyms.rst - source/_POST/DD/b/13/b13-trifecta-registry.rst - CLAUDE.md (especially Language Rules) - source/dict/framework.rst - source/_bib/references.bib (the PrototypeEvolvixCompiler entry) - source/matheology/hell/mm/b/19/b19-sgir_basic-gap-of-germs-2020-epidemiology-oov1_2026.rst (the b19 SGIR paper that uses Prototype Evolvix — claims here should match)

For each of the two hostile angles below, produce a numbered list of concrete holes with: (a) line number on the mirror page (or quoted sentence from the Zenodo deposit), (b) what the hostile reader would think and why, (c) proposed fix (replacement text or restructuring), (d) confidence in the criticism (HIGH / MEDIUM / LOW).

ANGLE 1 — Hostile reviewer for scientific reproducibility: - Strict standards for software-archive deposits. - Will check: is the version string traceable to a specific commit?

Are the binaries for each platform documented enough that a reproduction attempt could distinguish a real success from a false positive? Is the test methodology described in enough detail to be checked? Are claims about mathematical reliability supported by specific testable predictions?

  • Will fact-check claims like “given sufficiently large population sizes, the SSA and ODE timeseries should produce near-identical results, easily tested by anyone” — can a reader actually do this with the deposited binaries, given LLoL’s 2026 installation guide?

  • Will check whether the citation block makes the deposit findable and unambiguously citable.

ANGLE 2 — Hostile language-design expert: - Senior PL researcher familiar with DSLs for biology, ANTLR,

ProtocolBuffers, SUNDIALS, mass-action stochastic simulation, etc.

  • Will check whether the Implementation section is honest about the stack’s constraints (ANTLR-2-vs-3 era? muparser limitations? SUNDIALS version assumptions?).

  • Will challenge the “1:99 best-idea-rejection ratio” — is this a defensible empirical claim or a rhetorical flourish? Where is the evidence?

  • Will challenge the “Prototype” qualifier — does the page accurately disclose what is incomplete vs what is solid? Could a reader form an honest assessment of “should I use this for my research?”

  • Will spot any place where the page claims more than the binaries actually deliver.

ANGLE 3 (only if time permits) — Mirror-vs-Zenodo drift check: - Open the Zenodo deposit description text and the local mirror

page side-by-side.

  • Identify any substantive divergences (rephrased sentences, reorganized sections, changed wording on technical claims).

  • For each divergence: which is correct? The mirror or the deposit? (Per LLoL’s instruction, the Zenodo deposit is authoritative.)

  • Propose: revert the mirror to match the deposit, OR update the deposit to match the mirror, depending on which wording is better and which is canonical.

Output format: section per angle, numbered findings, then a synthesis section ranking the findings by importance for LLoL to address.

Constraint: cite lines verbatim before commenting. Do not fix the page in this session; produce a markdown report only. Mark suggested fixes as proposals, not applied changes.

Cross-references#

Created by: Claude Opus 4.7 Max in EDEN mode (2026m04d25), at LLoL’s request, paired with AA b13.