Sanitized reproducibility record
Hella Shield evidence index
The 41 artifacts selected for the V1 public evidence set, with artifact names and SHA-256 digests. Live lab results, offline analyses and formal models are kept visibly separate.
Evidence classes
PROVEN live: measured on owned disposable lab infrastructure.
PROVEN offline/formal: deterministic analysis, replay or state exploration over documented inputs.
OPEN: claims not promoted beyond the evidence currently available.
Core live evidence
| Exp | Claim | Artifact | SHA-256 |
|---|---|---|---|
| 001 | local process-control timing baseline | exp001-preliminary-batch.json | a8b06da53ff1af04468460d42242801ae171f916b2fd9b1754387a7076f16618 |
| 002 | live witness calibration and lower-layer overlap | exp002-live-witness-calibration.json | 29c93a78397546741b8740a02f8039181ba7e44291a6bb25f78cf249cb4e971d |
| 003 | live iSCSI filesystem-freeze containment | exp003-iscsi-freeze-10run.json | bdd4ba15cf49ffc4b841696f121c6ff127acf4c172b1e011915232e547dee92e |
| 007 | live pacing during a decision window | exp007-summary.json | 06ad7f042bf14f3cd4039bc51204c23a0de27eb50f048ef8388f8f0eca02da64 |
| 016 | first fresh blind policy falsification | exp016-summary.json | 549c4e591c1619a3105b7eb354784b34feb20c3b9c8c6877cf8228dd87d64175 |
| 017 | mutation-semantic blind challenge | exp017-summary.json | 553705939d18d808725d1d287668fb13d323b9a0ba195a4e857fd9a04455a3f4 |
| 018 | scoped change-intent blind challenge | exp018-summary.json | c96ff34dc326842f4afb8ce4dca461808bd2e889edb04685fe9e904837b44af9 |
| 019 | provenance + bounded-grace blind challenge | exp019-summary.json | 73ca3929c4afd8ec2b303a329e6044a1d1bcfaa7cc00dfb17a68e9a14c3b281c |
| 020 | risk-weighted mutation budget challenge | exp020-summary.json | 69f3e2c5d8346b005f5d16eebabe30e21f03682addb3aa8581ab68c0d28545c6 |
| 022 | live recovery-capability witness | exp022-live-result.json | a700298b9a52b7a406f78761da1de28ef43b51336d0faa84bf379df0837e877b |
| 023 | 10-run recovery-witness reproducibility | exp023-result.json | 360de94fc23c528b2d193a59993f1a39d93bc82694dc02fa7f2d06ff6bcf1ea1 |
| 025 | recovery proof while paced writes continue | exp025-result.json | 7a6907f07a14087b6eb975b2325d0397da47b0c28d33f23720c6ea5b41ce2080 |
| 026 | pacing rate vs proof-window exposure | exp026-result.json | b484f44a64d285a1300a158856b8315c699ab1a26bab021a72f8cf88e5eb761d |
| 028 | live credit-to-rate controller validation | exp028-live-controller-result.json | 9a2c4bbdc9bc4bcec4556606985193c140d6192834b94ee4177b67fe03b5527e |
| 029 | fail-closed finite-credit boundary | exp029-live-result.json | 9146d456aeff3bb5c174ca922c9fa139da4816cb4bb9469aeee4a6efde86e17f |
| 030 | credit-stretch live test | exp030-live-result.json | f4f32bbf2e5830b6ee26996d28f8460306a81f53c0ec668e85a8c17ceb589e5e |
| 031 | boundary-search live test | exp031-live-result.json | 27c7cb4636c44eb97fd0b75a4275ef9c3417071f52ae73e988e7df51ca45caa6 |
| 032 | byte-native governor live test | exp032-live-result.json | 2b38d22121f2243517ec56ca19848e8cc171a6e8f20d149374f144f4fc8fc801 |
| 034 | dual byte/time containment envelope | exp034-live-result.json | d6722e98e32b53a5422ed94c9146c0107292a8d1eb82a361908e7d27f16a235f |
| 035 | external supervisor of a non-cooperative writer | exp035-live-result.json | 86bdfa51dd238569a57b8f315580cd64a1deeb7f629f9e6cfbe1e76abc9862fd |
| 037 | multi-writer shared-credit overshoot | exp037-live-result.json | 1ab5023248122b5e30cb86815d22da8957944cad22514b045ea733abcea7230d |
| 038 | write-size sensitivity | exp038-live-result.json | 9dc7da814ca2208df6ce7fe8b1a04b361132d0b15393ecb923e1a25aaecb3254 |
| 039 | concurrency × write-size frontier | exp039-live-result.json | c52cb2f2ef1790e301fc93258e42312183bc3a7def697b6ad466f556148e25b5 |
| 043 | final combined preregistered V1 challenge | exp043-summary.json | d4dc404f4d00a7786715c6dafabf86fa4234c1d864a40623f9ec81c199d437c6 |
Formal and analytical evidence
| Exp | Claim | Artifact | SHA-256 |
|---|---|---|---|
| 004 | tamper-evident research receipts | exp004-receipts.json | 2c869231134576be5a53f2fda7cc90bdc295b3c728e01afca734b1256ac1a879 |
| 005 | recovery safety/freshness trade-off | exp005-result.json | a630f9b74b023f8834bf46da539f2d60e0e99a522840893a0f2a08614553c18c |
| 006 | staged-governor simulation | exp006-result.json | 86b2260a607e81811deb2bc6b739408a4735c42d51752cd10844f7d27a31c142 |
| 009 | witness-sufficiency ablation | exp009-result.json | 07b9e8cd580c18c248e95b9d3e6c8d1be630d1493d0bb33550a6edbb17f71efb |
| 010 | cost-aware containment frontier | exp010-result.json | 82562e200c08b698c4d046bc301d07b5e22096540a12db3f63716a50a6a02a9a |
| 011 | recovery-confidence v2 replay | exp011-result.json | de504642fc707ffee16868b3c7f48a82e4441a317c88c0e67dae0f342a78cf6d |
| 012 | selector perturbation stress | exp012-result.json | 1971b16cd65e545deaf22cd07f9115a03935d53aea191c65ed00c4e15b17f618 |
| 013 | abstention-to-pacing model | exp013-result.json | aa27414c5e100b030d81cfbf29d60deb36dc7a88a1be0d3adf25ab05e2585632 |
| 014 | temporal-semantic holdout replay | exp014-result.json | f01f3ddd09f3e132961f184c86175e5256e70fc7fbcc30cdb38ed7558d503264 |
| 015 | integrated policy composition replay | exp015-result.json | f101cb3ef130cdc6e9b1cecfabe35ddc619613c5db7e094b9e71446a2939b38e |
| 024 | recovery-backed write-credit replay | exp024-result.json | aad3dafb00d6e6c6c438aeda2b0de3b42b408ae902ba45f176265ee89ec04466 |
| 027 | credit-to-rate controller derivation | exp027-result.json | 6c090ff44b29af7b976119093ff533189ce8c446c272fa96de024fcad866bcb4 |
| 033 | credit-invariant verification | exp033-result.json | c18338c9e67b4984cf5d03175bdfc626fefd169c1b0b04006cb77712e5a80694 |
| 036 | three-clock containment analysis | exp036-result.json | 0dc6aaffa8e8d73da6710887227b90831191406fadf99babf0360f2f7d720f77 |
| 040 | reservation-before-admission finite-state model | exp040-result.json | 9941dc855fe9fd8be319032d812410a045a4b1863e1fde37f495baac2f396521 |
| 041 | shared-credit fairness model | exp041-result.json | 7456f60a29df583bebaf9299580a512087a352de3af6ac92c11a46702367d421 |
| 042 | evidence-loss authority-contraction model | exp042-result.json | 4d48347844aaaba417af7e922f49013daf1cb23e394d4287299d1dad210335c6 |
Verification manifest
All 41 publication artifacts can be checked against SHA256SUMS.txt. The manifest describes the sanitized publication copies; original local experiment evidence remains unchanged.
Truth boundary
The public record does not claim a live storage-side pre-admission implementation, enterprise-scale detector accuracy, universal ransomware coverage, or arbitrary application-consistent recovery. Those remain outside the proven V1 scope.
Public evidence is sanitized to omit credentials, private keys, long-lived tokens, private topology and operational details that are unnecessary to evaluate the research claim.