Only sourced facts: what each release changed, what the checks concluded, and which log it came from. Anything uncertain is marked [to be confirmed]; nothing is invented.
Data as of 2026-10-20
🟢 Verified — all gates passed🟡 In progress — underway, not done⚪ Planned — direction set, not started
v6Full-site relaunch2026-10-20
Verified
What changed
Full-site relaunch: the new site (multi-page, bilingual) was deployed and now serves vertciti.com; the three legacy deployments were deleted and the production source was replaced and committed. This relaunch follows the 2026-10-19 multi-page restructure recorded in the build-log stream.
New company-information and media-centre pages; English editions added for the data room and the AI knowledge base.
Contact form and subscription endpoints launched with the new site; console sign-in uses password plus authenticator-code two-factor.
Release check conclusions
Post-launch production audit closed: zero pages lost, security headers and four-viewport layout measurements all passed; open items (e.g., absolute hreflang URLs) are queued for remediation.
v3.2Evidence freshness mechanism2026-10-16
Verified
What changed
Evidence freshness mechanism shipped (spec cpr-freshness-mechanism-spec-2026-10-14): every evidence card's .data-src now carries a machine-readable data-asof attribute, so rendering and freshness checks read a single source. Cards not re-verified within 14 days automatically gain a pending re-verification badge and a dashed de-emphasis outline; past 30 days the number zone switches to a "verifying" placeholder instead of presenting stale figures as current. Historical archive cards (e.g. the 2026-09-27 incident review) are fixed records and are exempt.
Recompute-only, never re-date (this round's receipts): 5 life skills (counted in the skills directory), 10 activated teams (counted in the roster), 7 executives (counted in Organization public v2), 4 verifiable founder facts (checked one by one against public records), 0 named-evidence items (site-wide recount still zero) — five dates restamped to 2026-10-16 on recompute receipts, archived at deliverables/dispatch-2026-10-16/freshness-recalc-receipt-2026-10-16.md.
The 309 backend-test figure was deliberately NOT restamped: a full local re-run (488 tests: 459 passed / 25 failed / 4 errors) could not reproduce the original counting basis, so the figure and its 2026-10-15 date stand pending a counting-environment review; the freshness mechanism guarantees it surfaces automatically once stale.
Build log rolling: this entry is this week's real entry; build-log.json gained its 2026-10-16 entry (21 total). The freshness sentinel now judges card by card (website-freshness-check A11).
Release check conclusions
Three gates: data-room sensitive scan WARN (only the public contact email contact@vertciti.com, cleared for publication on review) / touch-target gate PASS (all 560 interactive elements >=44x44) / zh-en mirror parity PASS.
The new inline script's sha256 was synced into vercel.json; under the enforced production CSP, a real browser rendered both languages with zero script refusals, and aged simulations (26d / 45d) showed the badge and the verifying placeholder live, with the archive-card exemption behaving correctly.
v3.1BlackRock v3.12026-10-02
Verified
What changed
New numeric evidence bar under hero: 4 figures (309 / 5 / 10 / 7), each linking into the data-room; mobile gets its own 2-column layout; hero copy untouched.
Source + date labels next to every figure site-wide: evidence chips, archive card, 4 founder facts, roster-foot "data as of 2026-10-02"; codified as the .data-src component spec.
New footer honesty-boundary statement: three-state marker definitions + "not a commitment" wording.
Decorative-color usage boundaries on dark backgrounds are now documented; written clarification from the brand team is pending.
R10 exception wording changed to in-force status (brand team approved, now a brand constraint).
Nav lockup rounded to 120px; 8pt grid 4px sub-step documented as intentional.
390px touch targets: nav links / brand / expand-btn / link-arrow / cta-mail / footer links all ≥44px (19 substandard spots all fixed).
Release check conclusions
Automated checks passed: zero console errors, no overflow on either viewport, contrast and touch targets all compliant, zero external requests, reduced-motion mode verified.
CEO manual review passed (2026-10-03): no errors or overflow on either viewport; 19 substandard touch targets independently re-verified and fixed; evidence bar and source labels in place on both viewports; narrative intact.
v3BlackRock v3 · merged edition2026-10-01
Verified
What changed
Merged edition: v2 dark skeleton + E1–E10 animation system, content merges "v2 language + the live site story" — production-site render functions (founder / about / pipeline / skills / honest / governance / contact) quoted verbatim, no new facts.
7 section map: people / pipeline / five life skills / evidence / leadership & activated teams / letters / trust finale.
New E11–E13: evidence-chip stagger entrance, S4 card "in progress" breathing dot (#EAB308), top reading-progress bar (disabled under reduced-motion).
Honesty wording: removed donut + three-state distribution bar (demo data), rewrote chips to production-site honest wording, removed a fabricated "—chairman" signature, marked the S4 process metric [to be confirmed] (no figure yet).
Built with assertion-based string replacement; fixed 1 CSS nesting bug and removed 15 dead-CSS blocks during the build.
Gate results
Playwright real-render self-check passed: 0 console / 0px overflow / zero external requests; skill cards lit 5/5, evidence cards entered 3/3; all 5 contrast pairs ≥4.5:1; cross-language links work both ways.
Entered v3.1 as the "v3 approved build" (per build-log-v3.1 input declaration).
v2BlackRock v22026-10-01
Verified
What changed
First complete BlackRock institutional dark-language edition: 9-section structure + animation system; zero placeholders site-wide.
Photos unified to institutional duotone + vignette + grain + mono archive captions; no repeated outfit/scene in adjacent sections; standard white-background portrait never used as content illustration.
Two hands-on CEO edits: copy changed to public wording.
Gate results
CEO real-browser review PASS (2026-10-01), verdict "clear to show the chairman"; production untouched.
4 real-render combos (1440 / 390 / normal / reduced-motion): zero errors, zero overflow; contrast: spots switched to teal (7.55:1), all other teal-aa uses audited — no text on dark backgrounds.
E-1Changelog page (this page)2026-10-03
In progress
What changed
New changelog.html / changelog-en.html: reuse v3 CSS tokens (--ink / --slate / --paper / --line / --mute / --teal-aa) and embedded fonts (Space Grotesk / Inter), zero external requests; dark institutional feel, large type steps, data density.
Timeline only states facts found in the logs; anything uncertain is marked [to be confirmed] — none this round.
Footer nav of the v3 homepages (both languages) gains a "Changelog" entry linking here.
Release check conclusions
Automated checks passed: zero console errors, no overflow on either viewport, contrast and touch targets all compliant, zero external requests.
Independent review, brand check, and CEO manual review completed.
evidence archive
Evidence archive.
The homepage keeps only the strongest few items; the full evidence archive lives here: the number group, the founder's four items, the full evidence cards, and the named-evidence inclusion criteria. Raw records are reachable through the data room index.
StatusNumber group archiveAs of 2026-10-16
Homepage number group (full archive; the homepage shows one group)
671 · backend tests passing (671 collected · all passed) · Source: two identical runs of the counting-protocol script (miao repo e0fe1e2) · data as of 2026-10-19
5 · five life skills · Source: vertciti.com skills public records · data as of 2026-10-16
10 · activated teams · Source: data-room organization (public) v2 · data as of 2026-10-16
7 · executive layer · Source: data-room organization (public) v2 · data as of 2026-10-16
0 · named-evidence items · Status: 🟡 collecting · data as of 2026-10-16
FactsFacts & statements archive2026-10-16
The founder's 4 items (full-text archive)
Only human — All staff are AI; no human employees. The founder is the only human.
Look inward — We seek no transaction revenue from the outside; we seek your trust, earned against evidence.
No pretending — The company is in its building phase: no orders, no public cases. We invent no cases and write no revenue numbers.
Built in public — We take no outside commissions and sell nothing. The build is public in the Build Log.
Source note
Only human — off-site corroboration: the founder's real-name GitHub profile (@HUIIIM) publicly states "N=1 personal AI infrastructure" with public repository activity (a third-party platform trace); organizational composition follows Organization (public) v2 in the data room. Nature: self-statement plus platform trace, not independent third-party verification.
Look inward — a company policy statement; there is no external object to verify, and it is presented as a statement.
No pretending — a negative claim, directly checkable on this site's public pages (no case pages, no revenue figures); site-record basis.
Built in public — the build log is published on this site; public building activity is also visible in the public GitHub repositories (timestamps recorded by the platform).
2026-09-27 · A site-wide outage: what happened → what we learned → what we fixed
Source: build log entry · data as of 2026-09-27 (fixed record, exempt from freshness demotion)
What improved
Full site recovered; the lesson is now a release-gate iron rule.
What fell short
A page-script change shipped without syncing the security configuration; the browser blocked the main script and the page went blank — server status looked fine, so only a human opening the site caught it.
Lesson
After changing page scripts, security configuration must be updated in sync; post-release verification must open the site in a real browser — server status codes alone are not enough.
🟡 Collecting — The first verified named testimonial will grow evidence cards here. Now: 0.
Status: 🟡 collecting · data as of 2026-10-16
Inclusion criteria (full-text archive)
Verifiable real name — the speaker's real name must be verifiable; anonymous or pseudonymous entries are not accepted.
Verbatim quotes — no rewriting, no cherry-picked excerpts; long quotes fold open.
Verifiable source — every card must carry a clickable source link or an authorization archive ID; no source, no card.
Written consent — public display requires the speaker's written consent; scope and term are recorded at inclusion.
Withdrawable — speakers may request withdrawal or correction at any time; write to contact@vertciti.com.
Wording — under the N=1 model there is no paid relationship, so no commercial designation like "principal / contracting party" is used; we say correspondent / reader / public commenter.
Rules take effect first; the standalone page waits for the first 🟢 card.
This site sets no cookies and uses no third-party analytics. Only with your consent do we remember preferences locally in your browser and turn on performance measurement (Vercel Speed Insights).
