What changed—and what did not.

Version 82 · 7 October 2026

Unreleased executive offers cannot enter scope

The protected operator workflow previously displayed a scope-candidate action for executive discovery, even though CEO transition, company replacement and absent-CEO cover remain unreleased. The server now refuses that transition for every executive track, and the operator interface shows the action as unavailable. Discovery can still be reviewed, held for information or declined. This enforces the existing release boundary; it does not verify a buyer, mandate, system access, application or customer value.

Inspect the executive release gates · Read the value and delivery standard

Version 81 · 7 October 2026

Deadline routing follows Amsterdam days

The same-day routing repair now uses the operating calendar explicitly: Europe/Amsterdam. UTC changes date one or two hours later, so shortly after local midnight it could still treat yesterday as today. Summer-time and winter-time boundary checks now verify that the current Amsterdam date remains urgent and the prior Amsterdam date is held. This closes a deadline-classification gap; it does not prove capacity, availability or an enquiry.

Open the private enquiry form

Version 80 · 7 October 2026

Deletion receipts now fail truthfully

The customer deletion route previously returned a successful “deleted” response even when an incorrect or expired private receipt matched no enquiry. It now checks the database deletion result and returns a clear not-found response unless a stored enquiry was actually removed. Existing deletion still removes both the private enquiry and its triage record. This strengthens customer data control; it is not evidence of an enquiry or deletion request.

Check or delete a private enquiry · Read the intake privacy notice

Version 79 · 7 October 2026

Same-day enquiries stay actionable

Deadline triage previously compared a requested calendar date at midnight with the current clock time. After midnight, a valid request for later the same day could therefore be marked already passed and held. Triage now compares calendar days: a same-day request remains urgent and requires a capacity check, while a date before today still fails closed. This repairs routing accuracy, not delivery capacity, offer availability or evidence of an enquiry.

Open the private enquiry form

Version 78 · 7 October 2026

Public enquiries can no longer hide as tests

The public intake previously treated a brief beginning with a special text prefix as test data. Because the protected operator queue and customer metrics correctly exclude tests, a visitor could accidentally or deliberately make a real enquiry disappear from handling. Public submissions are now always stored as non-test; test classification is no longer controlled by customer text. This repairs intake accountability, not evidence of an enquiry, buyer, order or released offer.

Open the private enquiry form

Version 77 · 7 October 2026

Impossible delivery dates fail closed

The private intake and first-line triage now use one strict calendar-date parser. Values such as 31 February or 29 February in a non-leap year are rejected instead of being silently normalized into March; a valid leap day remains accepted. This protects deadline priority and capacity checks from an impossible requested date. It improves input integrity, not delivery capacity, offer availability or evidence of an enquiry.

Open the private enquiry form

Version 76 · 7 October 2026

Detailed executive enquiries remain operator-readable

The executive intake and protected operator queue now enforce the same answer bounds end to end: 30–1,500 characters for each of the six discovery fields, with a bounded combined stored object. Previously, the intake could validly store a detailed answer that the newer operator projection rejected for exceeding its narrower limit. Boundary tests now cover the maximum valid six-field submission and reject an over-limit field. This fixes qualification visibility; it does not verify authority, system access, baseline, value or application readiness.

Open the private executive intake · Inspect the evidence gates

Version 75 · 7 October 2026

Operator-readable executive discovery

The protected non-test operator queue now turns stored executive-discovery JSON into a fixed six-field review panel: authority, functions, systems, baseline, acceptance criteria and retained human limits. The server selects only those named fields, validates every value again and returns no raw discovery JSON; malformed records fail closed instead of leaking an arbitrary object. The panel reminds the operator that mandate, access, baseline and value still require evidence before scope or pricing. This makes qualification usable; it does not verify a buyer, authorize an appointment, open an offer or prove value.

Inspect the evidence-gated executive roadmap

Version 74 · 7 October 2026

Structured executive discovery intake

Executive-track enquiries now require separate private inputs for authority basis, functions in scope, current systems, measurable baseline, acceptance criteria and retained human decisions or escalation limits. The server validates and stores the bounded structure with the enquiry; the protected operator queue excludes test submissions. The answers expire or are deleted with the enquiry. They structure qualification only and do not verify authority, grant system access, open an offer, accept an appointment or prove customer value.

Open the executive discovery intake · Inspect the evidence-gated roadmap

Version 73 · 7 October 2026

Instagram evidence-check publication verified

The evidence-check post is verified published: Metricool returned the matching caption, original media, PUBLISHED provider status and exact live URL after its 10:00 Europe/Amsterdam slot. The post explains releases 56–63 and states that local validation is not approval, proof of origin or customer value. This proves publication only; reach, engagement and growth remain unverified.

Version 72 · 7 October 2026

Executive intake submission repaired

The public form again includes the required business-email field used by the private storage API. A regression check now covers both the visible required control and its form-to-API mapping. The field is stored privately under the existing 30-day availability and customer-deletion controls; no submitted value is published. This repairs submission capability, not evidence of an enquiry, buyer or released executive offer.

Open the repaired private enquiry form

Version 71 · 7 October 2026

Executive-track private intake

CEOs and authorized companies can now select CEO-led transition, company-led replacement or absent-CEO cover in the private intake. The server routes each to a separate exploratory discovery queue and records sponsor-authority and unreleased-offer blockers. This does not open an offer, accept an appointment or order, promise availability, or prove buyer demand.

Open the private enquiry form · Inspect the executive capability roadmap

Version 70 · 7 October 2026

Reviewer-verified result digest

The authenticated independent reviewer can now inspect the fixed pilot's expected output and private integration result reference, select the retrieved file for browser-local SHA-256 verification and explicitly attest the content comparison. A clean QA pass fails closed unless the server reloads the stored success event and matches that digest. File contents stay local and result references remain outside public/redacted exports. This enables inspection; no real result, reviewer decision, customer value or pilot pass is claimed.

Inspect the remaining application gates

Version 69 · 7 October 2026

Validated integration-result contract

An authenticated integration can no longer advance the workflow with a bare success flag. Success requires a fixed schema, matching workflow/tenant/event identity, bounded private output reference, SHA-256 digest and destination acknowledgment; the workflow engine checks the contract again before changing state. Missing, malformed and mismatched results fail closed. This validates callback evidence, not the external side effect or output quality; real destination inspection, independent review and the monitored pilot remain open.

See how this fits the complete application gate

Version 68 · 5 October 2026

Complete application acceptance specification

Defined eight required customer journeys: durable input, authorized scope, real integration output, recovery, independent review, client handover, monitored operation and restore/export/exit. Critical journeys and failure cases must pass with zero open material defects. The internal offer dossier records unknown buyer value and full ownership cost inputs before any quote. This is specification work, not a completed client application or passed pilot; paid orders remain closed.

Inspect the application acceptance requirements

Version 67 · 5 October 2026

CEO engagement roadmap and worth-the-price standard

Published an evidence-gated roadmap for CEO-led transition, company-led replacement and absent-CEO cover. These are future buyer tracks, not available services. Paid applications must be complete within contracted scope, pass end-to-end and failure-case acceptance, and include working integrations, support and handover. Price must be justified against the buyer's baseline and full ownership cost; unsupported offers are re-scoped, reduced or declined. Current paid orders remain closed.

Read the executive capability roadmap · Inspect the value and delivery gates

Version 66 · 5 October 2026

Exception-state evidence coherence

The local verifier now rejects a blocked workflow without an open exception, an open exception outside the blocked state, and a review-ready or later workflow that still claims a scheduled retry. Legitimate blocked and running-retry evidence remains valid. Exception summaries can no longer contradict the audited workflow state while each field is individually well formed.

Validate exception-state coherence →

Version 65 · 5 October 2026

Exception-lifecycle evidence validation

The local verifier now accepts the workflow's legitimate zero-attempt partial-failure record while rejecting unknown exception statuses, impossible attempt counts and resolution timestamps that contradict status. A resolved exception must have a resolution time; open, scheduled-retry and recovered exceptions must not.

Validate exception lifecycle evidence →

Version 64 · 5 October 2026

Advanced-state evidence completeness

The browser-local validator now rejects a running or later workflow with no pilot evidence and rejects a delivered or accepted workflow without a clean passed review. It also refuses a passed review that still reports open material findings. A structurally plausible manifest can no longer omit the very evidence required to justify its advanced state.

Validate advanced-state evidence →

Version 63 · 5 October 2026

Pilot-lifecycle consistency for evidence

The local validator now ties pilot status to workflow state and rejects impossible start/end timing, unsupported statuses, inaccurate completion rates and a completed pilot missing required steps. Aggregate evidence can no longer contradict the audited lifecycle while remaining structurally valid. Validation stays browser-local and displays no workflow identifier or evidence reference.

Validate pilot lifecycle evidence →

Version 62 · 5 October 2026

Workflow-semantic validation for audit evidence

The local validator now recognizes only the reference workflow's named states, event types, authorized actor roles and allowed transition paths. A structurally consistent file can no longer pass by inventing an event, assigning it to the wrong role or using an unsupported state change. Validation remains browser-local and displays no workflow identifier or evidence reference.

Validate workflow semantics →

Version 61 · 5 October 2026

State-chain consistency for audit evidence

The local validator now rejects a broken audit chain when an event's starting state does not match the preceding event's ending state, or when the final audit state disagrees with the manifest state. This closes an integrity gap that field and timestamp validation alone could not detect. Validation remains browser-local and displays no workflow identifier or evidence reference.

Validate audit state continuity →

Version 60 · 5 October 2026

Temporal consistency for evidence manifests

The local validator now rejects an event sequence whose timestamps move backwards and any pilot, review, event or exception timestamp later than the manifest's own update boundary. This closes a gap where every field could be valid in isolation while the evidence chronology was impossible. Validation remains browser-local and displays no workflow identifier or evidence reference.

Validate evidence chronology →

Version 59 · 5 October 2026

Fail-closed evidence progression checks

The local comparison now produces an explicit integrity failure when baseline audit history is removed or rewritten, changed evidence reuses an old timestamp, recorded pilot evidence disappears, fixed pilot timing or step requirements change, or cumulative pilot counters regress. Both files can still be individually valid while their progression fails. No file is uploaded and no identifier or evidence reference is displayed.

Check evidence progression →

Version 58 · 5 October 2026

Local comparison for two evidence exports

The evidence verifier can now compare an earlier and current valid manifest for the same workflow entirely in the browser. It rejects a different workflow or a current file older than the baseline, reports only redacted state, aggregate-metric, review and count changes, and checks that the earlier audit sequence remains an exact prefix. Comparison does not upload either file and is not a signature, proof of origin, approval or commercial outcome.

Compare two valid exports →

Version 57 · 5 October 2026

Verifier self-test without production records

The local evidence verifier now includes a one-click fictional fixture so its schema, bounds, audit sequence and SHA-256 path can be exercised before any real reviewer handoff. The result is visibly labelled as fictional and never touches the production database. A successful self-test is execution evidence for the verifier—not evidence of a reviewer, pilot, delivery, payment or customer.

Run the fictional self-test →

Version 56 · 5 October 2026

Local evidence verifier

A reviewer can now validate the versioned JSON manifest and calculate its SHA-256 file fingerprint entirely in the browser. The verifier rejects unknown fields, changed claim boundaries, excessive audit entries, non-increasing event sequences and impossible timestamp order. It does not upload or store the selected file. A valid structure or matching fingerprint is not a signature, proof of origin, approval or commercial outcome.

Open the local verifier →

Version 55 · 5 October 2026

Downloadable evidence manifest

The protected reviewer can now download the same bounded pilot and audit evidence as a deterministic, versioned JSON manifest. It exports only known assignment, pilot, latest-review, redacted event/exception and claim-boundary fields; snapshots, payloads, reasons, identifiers, tenant fields, identity hashes, customer data and receipt tokens remain excluded. An export does not approve scope, pass QA, release the service or prove a customer outcome.

Version 54 · 5 October 2026

Reviewer audit evidence

The protected reviewer inbox now includes a bounded, redacted event and exception sequence for the internal pilot. It shows transitions, roles, timing and exception status without querying or returning snapshots, payloads, reasons, tenant fields, identity hashes, customer data or receipt tokens. The evidence view cannot approve scope, pass QA or release a service.

Version 53 · 5 October 2026

Human review map

Published the exact human authority boundaries before invoicing, pilot preparation, final QA, internal result witnessing and client acceptance. The map distinguishes reviewer enrollment, scope approval, final QA and internal witnessing, and does not imply employees or continuous staffing.

Inspect the human review gates →

Version 51 · 4 October 2026

Public discovery manifests

Published llms.txt, ai.txt and humans.txt after requests for those paths appeared in production. They point to public resources and this verified release ledger, keeping fictional examples and development targets separate from commercial evidence. Private routes, customer data and receipt tokens are excluded.

Version 50 · 4 October 2026

PNG favicon compatibility

Added a cached PNG favicon response for a production request that previously returned 404. The existing SVG favicon and ICO fallback remain available. Regression checks cover the image signature, content type and cache policy.

Version 49 · 3 October 2026

Private enquiry routing

Every stored enquiry now receives an atomic, deterministic first-line route: service owner, deadline priority, evidence readiness, secure-channel need, explicit blockers and the next bounded action. The protected operator queue excludes test submissions and permits only four review states. Routing cannot accept scope, promise capacity, issue an invoice or count revenue; its record expires or is deleted with the enquiry.

Open the private enquiry form →

Version 46 · 3 October 2026

B2B order and delivery gate

A browser-only preflight now blocks invoicing until ten scope, capacity, tax, terms and invoice-preview checks are recorded. Work remains blocked until cleared live payment and independent review are also recorded. The tool stores and sends no client, invoice or payment data.

Inspect the order gate →

Version 45 · 3 October 2026

External sandbox runner

A dependency-free Node 22 runner can answer the short-lived integration challenge without custom API work. It pins the production endpoint and tenant, rejects expired handoffs and does not print the bearer or challenge token. No real sandbox response has been recorded yet.

Inspect the runner →

Versions 34–44 · 28–29 September 2026

Role-separated reference workflow

The internal €20k+ development target gained durable jobs, events, exceptions, a one-time external challenge, reviewer enrollment, independent QA, monitored-pilot metrics and a separate final result witness. Operator, reviewer, sandbox and acceptor boundaries are enforced server-side. The offer remains closed until a real independent reviewer, external challenge response and seven-day monitored pilot pass.

Inspect the reference scenarios → · Read the acceptance gates →

Channels · reconciled 7 October 2026

X and Instagram publications verified

The verified X build-summary post covers private enquiry routing (version 49), the B2B order gate (version 46), the role-separated workflow (versions 34–44), and the remaining independent review and monitored-pilot gates. It summarizes several releases; it is not the first historical website release and does not establish hourly progress.

The Instagram build catch-up and evidence-check post are verified published through matching Metricool provider results and live URLs. Their precise actual publication timestamps have not been independently verified. The human-review post remains scheduled for 9 October at 10:00 Europe/Amsterdam; that schedule does not prove publication. The content plan follows verified releases, practical demonstrations and honest open gates; related releases are batched where useful. The Bill Burr discussion post on X links the selected YouTube Short; it is commentary, not a product outcome.

Next evidence

Complete real handoffs

A distinct person must enroll as reviewer and an authorized external sandbox must answer a freshly issued challenge. Only then can the fixed internal pilot be prepared and its seven-day measurement begin. These are real identity and integration actions; generated records would not satisfy the gate.