Conformance record
| Field | Value |
|---|---|
| Applicable profile | Evidence-linked topic wiki using a targeted evidence map and thematic case synthesis |
| Highest implemented level | Level 4 Graph-Linked; no Level 5 or Level 6 claim |
| Renderer | Zensical 0.0.50 stock-default UI, production |
| Adapter | Isolated, user-switchable desktop width override; repository-graph and Ask Wiki component assets are otherwise content-scoped |
| Sunset UI | Frozen non-production profile (zensical.sunset.toml, assets/theme.css) |
| Material | Omitted by explicit user choice |
| Authority chain | Primary source → report → claim → mechanism → candidate requirement → assurance test; statement type and transferability limits remain explicit |
| Evidence status | Current pass: 165 retained candidates, 112 fully extracted included reports, 53 exclusions; complete 379-row 2018 publication-cohort enumeration/extraction with 36 retained cohort decisions and 343 rows lacking contextual dispositions; targeted 2019–2026 discovery, a reproducible 2025–2026 publication-date update screen and limited later-migration checking; official source URL retained for every included row; not exhaustive interval-wide and not a prevalence estimate |
| Generated evidence views | Complete 112-report catalogue, corpus profile, selection/coverage flow, assurance/uncertainty dashboard, ten-record counter-position register, individual report pages, per-mechanism membership lists, case–mechanism/evidence map, official-response catalogue and data/citation checksums generated from canonical records, with freshness, lineage and row/page-count checks |
| Defence traceability | Stable candidate REQ- and designed AST- registers plus a generated finding-section crosswalk with compact route summary; Defence validation and test execution are not claimed |
| Repository graph | One complete enriched Graphify dataset drives searchable/filterable 2D and true force-directed 3D views, with named communities, six technical interconnection layers, self-ingestion prevention, exact data-parity/integrity checks and a responsive control panel; demoted under About and tools and not study evidence |
| Generated discovery | Content-aware llms.txt, llms-full.txt, and llms.json with authority metadata and freshness check |
| OKF sidecar | Not implemented; no Level 6 claim |
| Browser evidence | Current stock-default profile pass covered the home route, complete report catalogue, a 2018 generated report, findings, Defence requirements/tests, the 2018 method audit, search, two-level nested navigation, mobile drawer, narrow tables, visible keyboard focus, Mermaid rendering and both repository graphs. The 17 July width-switch pass additionally covered Standard → Wide → reload persistence → Standard at 1404×647, exact site-title/article and Search/page-outline text-rail alignment in both desktop modes, and the hidden control/native grid at 390×844. The 20 July Ask Wiki pass covered desktop and narrow form layout, correct /search.json retrieval, one-result-per-page diversity, missing-password validation, fail-closed endpoint behaviour, six retrieved links and console errors. The 21 July access-control pass confirmed the password field uses autocomplete=off, is cleared after a submission, creates no authenticated browser state, and reports the static local server's malformed API fallback as an error. The 22 July Executive summary pass covered the grouped Home navigation, active summary route, page outline, warning, summary table and mobile drawer at desktop and 390×844. The 23 July Graphify pass covered the official-response catalogue, initial crosswalk, glossary, output-status table, grouped navigation and both complete Graphify views. The subsequent maturity pass covered the corpus profile, assurance dashboard, counter-position register, selection-flow Mermaid, report-level counter-position, revised Defence crosswalk and nested drawer at desktop and a 433px effective narrow viewport. The narrow selection diagram and three-column crosswalk remained within document width; tables retain local overflow where required. No console errors or page-level horizontal overflow were observed; broader screen-reader testing remains open. No Level 5 claim. |
| Full runtime/release discipline | Not claimed |
| Local runtime | LaunchAgent-backed regulation-28-wiki service at http://127.0.0.1:8000/; registered, loaded, HTTP 200, restart tested |
| Last verification | 2026-07-23 maturity pass: canonical regeneration; 165/112/53 screening reconciliation; 379-row 2018 coverage overlay with 343 undispositioned rows exposed; 112 unique primary-source/report joins; 107 linked response letters across 63 reports; ten source-resolved DIS- records; generated corpus, assurance, response, counter-position, data/citation and finding/requirement/test routes; content, claim, requirement/test, abstract, Ask Wiki, profile-isolation, discovery-freshness and Graphify parity/integrity checks; strict production and sunset builds. The rebuilt search index contains no legacy compatibility records and the always-on profile returned HTTP 200. Pages rejects the Workers ratelimits binding, so no per-user or WAF rate limit is claimed. |
Production intentionally exposes the native Zensical baseline. The reader model now separates evidence, findings, Defence application, methods and outputs from operational tooling. The repository graph remains a Graph-Linked implementation contribution, but the generated case–mechanism/evidence map is the reader-facing analytical map. The sunset UI is retained for comparison, not as a supported adapter.