From 552672addea22939491af49585cfc120a53add3d Mon Sep 17 00:00:00 2001 From: Paul Clark Date: Sun, 25 Oct 2026 12:08:35 -0400 Subject: [PATCH] Record paper-review Hub publication evidence --- docs/encyclopedia-roadmap.md | 3 ++- docs/paper-review-extensions.md | 45 ++++++++++++++++++++++++++++++--- 2 files changed, 43 insertions(+), 5 deletions(-) diff --git a/docs/encyclopedia-roadmap.md b/docs/encyclopedia-roadmap.md index a91aafa..852a126 100644 --- a/docs/encyclopedia-roadmap.md +++ b/docs/encyclopedia-roadmap.md @@ -441,7 +441,7 @@ Phases express a sensible teaching/build sequence, not promised dates. Recheck t These source-informed gaps were confirmed against fresh repository main branches on October 24, 2026. They are retained as **Planned, unprioritized follow-ups**: they do not change the accepted proposal queue, the 48 accepted encyclopedia proposal IDs or the phase counts. `RR-*` is a local reading-review identifier, not an encyclopedia-audit ID. Existing mechanics named in each row must be reused rather than rebuilt. -The separate [paper-review implementation checklist](paper-review-extensions.md) records six 2026-10-25 paper-derived decisions without changing this queue or denominator. PR-03's bounded weak/repaired Groth16 relation, PR-05's bounded Off-Key adaptation, PR-06's exact abort/retry lesson and PR-10's quantitative FRI correction are live. PR-04 remains explicitly deferred for missing CKKS public-key/CoeffToSlot/Galois-rotation prerequisites; the lab's existing relinearization key switch is not that missing path. PR-11 is planned and contributes prior-aware controls to RR-POWER-MEASURED without adding a deep-learning stack. That checklist owns paper versions, reviewed sections, dependencies and delivery evidence so they are not duplicated here. +The separate [paper-review implementation checklist](paper-review-extensions.md) records six 2026-10-25 paper-derived decisions without changing this queue or denominator. PR-03's bounded weak/repaired Groth16 relation, PR-05's bounded Off-Key adaptation, PR-06's exact abort/retry lesson and PR-10's quantitative FRI correction are live. PR-04 remains explicitly deferred for missing CKKS public-key/CoeffToSlot/Galois-rotation prerequisites; the lab's existing relinearization key switch is not that missing path. PR-11 is planned and contributes prior-aware controls to RR-POWER-MEASURED without adding a deep-learning stack. [Hub PR #291](https://github.com/systemslibrarian/crypto-lab/pull/291) publishes their exact reviewed pins, exhibit maps and corrected catalog scope. That checklist owns paper versions, reviewed sections, dependencies and full delivery evidence so they are not duplicated here. | ID | State and target | Confirmed gap and bounded acceptance criteria | | --- | --- | --- | @@ -1795,6 +1795,7 @@ For each proposal, append a dated entry below when its state changes. Preserve t | 2026-10-25 | PR-10 FRI quantitative-claim correction | [STARK Tower PR #23](https://github.com/systemslibrarian/crypto-lab-stark-tower/pull/23) merged as exact main `51b29f35a0c5`, with reviewed feature and merge trees identical. Final-head run 40203974891 and exact-main/deployment run 40204331185 passed the complete computational, DOM and claim gates, all four focused mutations from a passing baseline, publishing/maintenance/build checks and all 24 hosted browser/accessibility tests. Pages artifact `12602817749` identifies the exact merge; all three public files matched it byte for byte. The live ledger separates exact arithmetic, conditional theorem terms, heuristics and empirical fixture outcomes; fails closed on missing proximity, fold-MCA and end-to-end protocol terms; caps union bounds; rejects invalid domains; and does not call sampled rejection a universal soundness proof. PR-10 is **Live** and does not change the encyclopedia proposal count. | | 2026-10-25 | PR-05 Off-Key selective-failure delivery | [SPDZ Forge PR #27](https://github.com/systemslibrarian/crypto-lab-spdz-forge/pull/27) merged as exact main `2bfdc83b0ef5`, with reviewed feature and merge trees identical. An independent CL079 paper-to-code audit re-derived the consistent `Delta(g-d)` and inconsistent `Delta(g-y)` residuals, unmasked recovery, masked bit-promise inference and attacker/omniscient boundary without finding a defect. Final-head run 40204049873 and exact-main/deployment run 40204752389 passed 76 unit tests, all three focused mutations from a passing baseline with sources restored, six maintenance checks, the production build and all 18 hosted browser/accessibility tests; exact main reported zero dependency vulnerabilities. Pages artifact `12603930373` identifies the exact merge, and all three public files matched it byte for byte. The live lesson keeps ordinary corruption, legitimate additive MAC-key shares, inconsistent reconstructed contexts, privacy leakage and its bounded three-party trusted-dealer adaptation distinct. PR-05 is **Live** and does not change the encyclopedia proposal count. | | 2026-10-25 | PR-03 unintended-witness and constraint-repair delivery | [SNARK Arena PR #26](https://github.com/systemslibrarian/crypto-lab-snark-arena/pull/26) merged as exact main `025f878b22b2`; reviewed feature and merge trees are identical. Final-head run 40207370988 and exact-main/deployment run 40207754121 passed 27 artifact-provenance controls, 51 unit tests, eight maintenance checks, typecheck/build, all 42 hosted browser/claim/accessibility tests and the clean-baseline missing-constraint mutation; source and bundle were restored. Provenance artifact `12604231532` records reproduced-and-bound circuit material, and Pages artifact `12604427228` identifies the exact merge. An independent exact-main build, the Pages artifact and all 18 public files matched byte for byte. The live lesson proves an explicit unintended `(2,2,r-4)` assignment against the weak Groth16 relation, rejects it with separately regenerated repaired-relation artifacts, retains all four legitimate binary witnesses, and keeps circuit repair, client validation and ordinary proof tampering distinct. PR-03 is **Live** and does not change the encyclopedia proposal count. | +| 2026-10-25 | Paper-review, A-D-12 and L-R-01 Hub publication | [Crypto Lab PR #291](https://github.com/systemslibrarian/crypto-lab/pull/291) merged as exact main `2927f2a2b9a0`. Seven PR workflows passed all 13 applicable jobs and seven exact-main workflows passed all 14 jobs, with only the event-inapplicable drift job skipped in each event. Pages run 40209975187 deployed artifact `12604959808` through successful deployment `7315606531`. The public catalog and knowledge-map JSON matched the artifact and exact main byte for byte. The Hub now publishes exact reviewed pins and exhibit maps for PR-03, PR-04, PR-05, PR-06, PR-10 and PR-11, including the honest CKKS deferral and planned Power Trace controls, while also publishing the already-live A-D-12 and L-R-01 catalog evidence. Full paper versions, limits, workflow links, hashes and unrelated stale-pin boundaries remain in the paper-review checklist. | Apply the repository’s existing workflows from [CLAUDE.md](../CLAUDE.md) and the relevant lab’s instructions. Re-read current source before upgrading a pin. Generated algorithm fields and catalog/README output remain generator-owned. Preserve exact evidence and current-source guards; do not turn a README mention, UI label or failed finite search into an implementation/security claim. Each rendered result must be derived from that run, and changed inputs must retire stale verdicts. diff --git a/docs/paper-review-extensions.md b/docs/paper-review-extensions.md index 55b7251..ea55161 100644 --- a/docs/paper-review-extensions.md +++ b/docs/paper-review-extensions.md @@ -336,8 +336,9 @@ changed merely to make the paper visible. #### Catalog corrections found during reconciliation -The current central-catalog change set stages the corrections found during this -reconciliation: CKKS Lab no longer advertises a reproduced chosen-plaintext or +[Crypto Lab PR #291](https://github.com/systemslibrarian/crypto-lab/pull/291) +delivered the corrections found during this reconciliation: CKKS Lab no longer +advertises a reproduced chosen-plaintext or key-recovery attack; Power Trace's key-recovery credit points to its executable bounded recovery and its unimplemented NTT/Shamir references are removed; SNARK Arena removes false discrete-log and PLONK implementation credit while preserving @@ -345,8 +346,44 @@ the toxic-waste forgery in learner copy and its pinned exhibit (the attack vocabulary has no honest term for that failure); SPDZ Forge describes BigInt field arithmetic with WebCrypto SHA-256 commitments; and Proof Tally attributes its local Field64/Prio3 and exact rational DP logic separately from noble's -TurboSHAKE128. These are staged until the Hub change itself passes, merges and is -verified live. +TurboSHAKE128. The same Hub delivery publishes exact reviewed pins and exhibit +maps for all six targets without changing the recommendations' dispositions. + +#### Central catalog delivery + +PR #291 merged as exact main +[`2927f2a2b9a0`](https://github.com/systemslibrarian/crypto-lab/tree/2927f2a2b9a08dc7d5d1598e38bb60be10fb5709). +Its seven PR workflows passed all 13 applicable jobs, with only the +event-inapplicable drift job skipped: [README/catalog](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640766), +[Teach pages](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640791), +[learning games](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640843), +[public evidence](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640876), +[protocol workshops](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640757), +[arcade games](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640712) +and [maintenance verification](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209640833). +The logs report zero failed tests, 24 live worksheets with 554 controls and zero +drift, and successful final maintenance stages. The six target review pins match +their current clean mains; 51 mapped exhibits and every recorded target anchor +validate. Fifteen separately reported full-fleet stale pins belong to other labs +whose mains moved and were not silently re-reviewed here. + +Seven exact-main workflows then passed 14 jobs with the same single skip: +[README/catalog](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975632), +[Teach pages](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975673), +[learning games](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975631), +[public evidence](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975608), +[protocol workshops](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975602), +[arcade games](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975679) +and [Pages](https://github.com/systemslibrarian/crypto-lab/actions/runs/40209975187). +Pages artifact `12604959808` (5,748,550 bytes, SHA-256 +`785c40b07639a0f7fe0f92e7432ddb75cc65b9b41feef693e5dae0d2da803ee8`) +produced successful deployment `7315606531`. The public catalog (821,746 bytes, +SHA-256 `4338aeb90432f4a2ae1a327cde45146c5634dfaf108fd157d1e024115e7c2125`) +and knowledge-map JSON (257,251 bytes, SHA-256 +`b893d0099741da57825baea0cedb9d7095d6d3e657645b6b705bf25d1422140c`) +matched the artifact and exact main byte for byte. Served `README.md`, `CATALOG.md`, +the encyclopedia roadmap and this checklist before the present evidence-only +follow-up matched exact main as well. Across the initial PR-01, PR-02, PR-07, PR-08 and PR-09 deliveries and six reconciled recommendations, unrelated cryptographic engines and fleet