Material changes should be inspectable.

The Readiness Brief: a daily blog, new every morning at 10:00 ET

• New Blog section with a daily brief on clinical data readiness. Each weekday has its own format: Standards Decoded, Lock Room, Reviewer's Eye, Capacity Math, Failure Friday, a Weekend Deep Read and a Sunday Week Ahead.
• Every brief follows one pattern: a short version, the explanation with sources, a synthetic example, three things to do today, one question to check yourself and links to the matching Navigator tool or Handbook chapter.
• Briefs appear at 10:00 Eastern on their date, adjusting for daylight saving time, without a site rebuild. An RSS feed, a daily calendar reminder and a reading streak kept on your device help you follow along. The home page shows today's brief.
• Blog added to the main navigation. No change to existing pages, tools, themes or colours.

Practical Leadership Handbook, simpler navigation and a two-path landing page

• New Handbook, Clinical Statistical Programming Operations: A Practical Leadership Handbook, built into the platform. The twelve-chapter structure is in place. Three chapters are written in full as author drafts: Resource and Capacity Planning (4), Dry Run and Database Lock Readiness (6), and Dependency, Risk and Escalation Management (8). The other nine are shown as clearly labelled outlines, not as finished text.
• Every written chapter follows one pattern (operational problem, practical methodology, realistic example, common failure modes, recommended actions), then a downloadable CSV template, a link to the matching Navigator tool, numbered references and its review status. Worked examples use a fictional study and the Capacity Estimator’s own illustrative defaults; they are not benchmarks or case studies.
• Handbook full-text search across all twelve chapters, a references page, and a “My reading” panel that remembers progress on your device only. No login, no paywall.
• New “Review the Handbook” page: an invitation to review a chapter in about 45 minutes against a synthetic scenario. It shows the review status plainly: no independent review has been completed yet.
• Navigation reduced to five items (Navigator, Workspace, Handbook, Resources, About) with section tabs. Intelligence, Simulator, Capacity Estimator, Playbooks and the AI workflow builder now sit under Navigator; Reminder Center under Workspace; Field Guide, Standards watch, Regulatory scenarios, Practice Exchange, Glossary and the walkthrough under Resources; Framework, Methodology, Validation, Review panel and this changelog under About.
• New home page with two paths, Build a Readiness Plan and Explore the Handbook, plus a start-from-your-situation row. Tool pages gained “Learn the method” links to the matching chapters.
• AI Prompt Studio is now the AI workflow builder (page address unchanged). The describe-your-challenge builder comes first; the twelve ready-made workflows, weekly rhythm and prompt shelf sit in one collapsed block that opens when you link to it. Behaviour and the live-AI safeguards are unchanged.
• Resources reorganised as a library with the three Handbook templates; the duplicate regulatory-scenarios panel was removed. Privacy page now describes Handbook reading, search and what is counted. About and the review panel describe the Handbook and its review status.
• New tools/site_chrome.py and tools/apply_chrome.py: one source for the header, section tabs, footer and “Learn the method” strip on every page, with a --check mode. New tools/build_handbook.py and handbook-src/: chapters are Markdown source files that generate the static Handbook pages and the search index, and the build fails if a written chapter breaks the common pattern or cites a reference that does not exist. The deployed site still needs no build step.
• New assets/css/v34.css, assets/css/handbook.css, assets/js/handbook.js and assets/js/ai-builder-page.js. New aggregate event names (chapter opened, template downloaded, tool opened, search found something) carry no typed text. Version strings set to 34. Footer simplified to four columns.
• Removed an internal evidence-planning term from the public v30 changelog entry.

Prompt Studio: describe a challenge in plain words, get a start-to-finish workflow

• New “Describe your challenge” builder at the top of Prompt Studio. Type what you are stuck on and get a numbered workflow with an owner, inputs, outputs, a time estimate and review checkpoints for each step, plus assumptions, risks, what to verify and what “done” looks like. Copy it as text or Markdown, download it as Markdown or JSON, print it, tick steps off, or save it to your shelf. No prompt to write, copy or paste.
• The builder uses free AI models only, with an automatic fallback to the next model when one is rate-limited or unavailable. Free models have strict rate limits, so you may be asked to wait, and the daily allowance can run out. The page says so plainly and offers the closest of the twelve ready-made workflows when live AI is busy, off or unavailable.
• Confidentiality: a prominent warning states that free-model providers may log prompts, that the text leaves your browser, and that only synthetic or generalized text should be entered. A confirmation box is required, and sending is blocked (in the browser and again on the server) when text looks like an identifier, email address, phone number, confidentiality marking or key. The check is a heuristic, not a guarantee.
• Return-visit features kept: today’s pick, last-7-days tracker, weekly reminder, prompt shelf (now also holds saved AI workflows) and the twelve ready-made workflows, which still work offline and without any AI service. New: a list of your recent AI workflows with step progress, kept on your device only.
• Privacy page now describes the optional live builder, what leaves the browser, and what the server does and does not keep.
• New serverless function api/workflow.js (logic in lib/studio-ai-core.js) that calls OpenRouter. The API key lives only in the OPENROUTER_API_KEY environment variable. Only model ids ending in :free can ever be called: the list is fixed in code, the browser cannot choose a model, the request body is built from an allowlist, and a final guard refuses any non-free id before a request is made. Retry with backoff, per-attempt and total time limits, per-visitor rate limits (4 a minute, 20 a day, per instance), an instance-wide cap on upstream calls, an input length cap (1,200 characters) and same-origin enforcement with an optional origin allowlist.
• New vercel.json, package.json, .env.example, .gitignore; studioAiEndpoint setting in assets/js/config.js; shared assets/js/studio-workflow.js. 54 new dependency-free tests against a mocked OpenRouter (tests/ai-proxy-tests.js, also runnable in a browser via tests/run-in-browser.html); the 33 Studio and 29 engine tests still pass. Version strings updated.

AI Prompt Studio: twelve reusable, confidentiality-aware readiness workflows

• New AI Prompt Studio page with twelve fill-the-blanks workflows: think through a readiness problem, run a pre-mortem, prepare for a cross-functional meeting, run a weekly readiness review, turn an experience into a shareable lesson, turn study files into an outcome, write a stakeholder status update, build a tiny offline tool, turn manual steps into agents with human review checkpoints, check an AI draft, pressure-test an answer, and practise with a quiz drill on any Field Guide topic. Each prompt asks the assistant to collect context through tap-to-answer questions first, then deliver the result directly.
• Reasons to return: a daily pick, a last-7-days tracker, a recurring weekly-review calendar reminder, a weekly review that can read your open Workspace actions, and a browser-local prompt shelf you can export and import.
• Synthetic sample pack for the fictional study ZEPHYR-101 (findings log, programming tracker, vendor transfer log, team notes, late change request) for practising without real data.
• A confidentiality check warns when what you typed appears to contain identifiers, emails, confidentiality markings or secrets. It is a heuristic reminder, not a guarantee.
• Every operational playbook and Readiness Intelligence result links to a prefilled “Think it through with AI” prompt; every Field Guide chapter links to a practice quiz.
• New nav and footer links on every page; version strings updated; 33 dependency-free tests for the Studio engine. Prompt Studio stores drafts, shelf and usage history only in browser storage and sends nothing to a server.

Scenario Simulator, Capacity Estimator, and validation protocol

• New Scenario Simulator: test a database lock moved forward, programmers unavailable before an interim analysis, or a safety-update workload. Deterministic rules show the critical chain, fixed waits that staffing cannot shorten, QC independence limits and a recovery option, with the rule ID behind every finding.
• New Capacity Estimator: P10 to P90 effort and weeks with QC and review included, calibration from your own completed-work actuals (processed in the browser), and exportable assumptions, exclusions and change-order triggers. Defaults are illustrative, not benchmarks.
• New Validation Protocol page: a pre-specified, draft study design for testing both tools against independent practitioner judgment. No results exist yet; the tools are expert-derived and not validated.
• New unit tests for both engines (29 tests: schedule arithmetic, monotonic properties, reproducibility, calibration).
• Housekeeping: version strings, citation, README and quality report brought up to date; web app manifest added; new nav and footer links on every page.

Independent validation follow-through: adaptive plans, cross-study conflicts, and evidence linking

• Navigator plans can now carry study-specific known risks and blockers: each one you log is woven into the relevant checkpoint and tracked until you mark it resolved, instead of living only in free-text context.
• Workspace now flags resourcing conflicts across saved plans that name the same shared person, role, or pool with critical work due within days of each other.
• The Workspace readiness signal (ON TRACK / AT RISK / NOT READY / READY) now has an on-page explainer of exactly what drives it, and now also accounts for open plan-level blockers, not only task status and dates.
• Each generated plan now links directly to its matching downloadable evidence tracker in Resources, where one exists, instead of leaving the two disconnected.
• "Save to Workspace" is now visually separated from secondary plan actions (calendar, copy, print) and shows an explicit saved/unsaved status.
• This changelog is now split into Guidance & practitioner-facing changes versus Engineering & maintenance changes, with a filter above.

Dropdown behavior, audio reliability, and versioning policy

• Reverted the brief homepage reordering introduced right after v31.0 per site owner feedback; the homepage returns to its v31.0 section order.
• Readiness Intelligence's Signal, Region/authority, Milestone, and Accountability lens controls no longer load with a value pre-selected, and no longer show results until the visitor fills in every field and selects "Analyze impact" — consistent with the same principle already applied to Navigator's role selection in v25.0.
• Fixed the walkthrough's "Play narrated walkthrough" control sometimes being stuck disabled due to Chrome/Windows loading available voices asynchronously; it now stays usable and retries voice detection instead.
• Adopted an immutable-release policy: each future update is published as its own new version folder rather than overwritten in place.

Independent audit response

Produced from a full technical, UX, accessibility, product, and evidence-governance audit of Version 30.0.

• Unified primary navigation and the footer across every page; several pages were missing links (including Walkthrough) or marking the wrong page as current.
• Fixed a duplicate page title, a skipped heading level on five pages, and two failing text-contrast pairs.
• Added a lightweight rate limiter to the feedback and analytics endpoints, and fixed an analytics event that was tracked client-side but silently rejected by the server.
• Closed a localStorage privacy gap where a Practice Exchange draft could retain a contributor's name indefinitely after submission.
• Hardened Workspace and Reminder Center against malformed saved data, and fixed a blocked-popup and an anti-spam edge case with no user feedback.
• Wired up a previously dead Navigator feature (role-shortcut buttons) and removed one that was never reachable.
• Softened "cross-functional" and "monitoring" language in the homepage, trust band, and Standards Watch copy to match what the product's own Content Validation Status table discloses.

Recurring-use and practitioner-review release

• Added Standards & Readiness Watch with curated authoritative-source entry points, browser-local review status, filtering, and a monthly calendar reminder.
• Added Reminder Center to turn open local Workspace actions into a filterable agenda and calendar export.
• Added eight searchable Operational Playbooks with trigger, first actions, handoffs, closure evidence, and Navigator links.
• Added copy, text-download, and print/PDF briefing exports to Readiness Intelligence.
• Added a Practitioner Review Panel pathway with confidentiality-safe review boundaries and scope-specific evidence documentation.
• Expanded privacy-conscious aggregate measurement and updated release, citation, sitemap, accessibility, and administration documentation.

Readiness Intelligence

• Added a Change-to-Action Analyzer for five common operational signals: standards alignment, late analysis change, health-authority request, external-data delay, and validation finding.
• Maps each signal to affected artifacts and functions, prioritized actions, timing, risk controls, and closure evidence.
• Added a direct, prefilled handoff from Readiness Intelligence into Navigator for dated role-specific planning.
• Extended the CDOR method with Signal → Applicability Decision → Affected Artifacts → Controlled Action → Closure Evidence.
• Preserved the v28.1 Navigator, Workspace, Field Guide, resources, walkthrough, privacy, and evidence-governance capabilities.

Walkthrough refinement

• Added clickable chapters and constrained optional narration to suitable natural English device voices while retaining embedded visual captions.

Guided walkthrough

• Added the 56-second visual product walkthrough, transcript, and discoverable entry points.

Orchestration typography refinement

• Increased the size and contrast of micro-labels and supporting text inside the Clinical Data Operational Readiness instrument panel.
• Promoted the closing line “Make invisible readiness gaps visible before the milestone does.” to a clearly readable concluding statement while preserving the existing panel geometry.

Readiness management and replanning

• Replaced binary task checkboxes with six operational statuses: Not started, In progress, Complete, Blocked, Deferred, and Not applicable.
• Added dependency-aware soft gates when users move work forward out of sequence, with documented override rationale instead of rigid workflow locking.
• Separated completion percentage from the CDRN readiness signal, including ON TRACK, AT RISK, NOT READY, and READY states based on blockers, overdue critical actions, dependency gates, and milestone timing.
• Added overdue indicators, editable action dates, milestone replanning, impact preview, preserved completed-action dates, and local change history.
• Upgraded collapsed Workspace summaries with status counts, next action, target history, completion, and readiness state.
• Redesigned the Clinical Data Operational Readiness signature visual as a premium orchestration instrument panel with clearer context, control, and output layers.

Workspace portfolio view and neutral role selection

• Rebuilt Workspace as a compact project accordion so multiple saved plans remain visible while only one is expanded at a time.
• Added completion and next-open-action summaries to each collapsed project row.
• Removed the default Statistical Programmer selection from Navigator; users now explicitly choose the accountable role.
• Rebalanced the CDOR orchestration visual to keep Evidence controls and Evidence checkpoints distinct and added subtle categorical styling.

Luxury-tier visual refinement

• Introduced a refined charcoal, warm-porcelain, teal, and restrained champagne visual system across the full site.
• Upgraded the header, page heroes, signature readiness-plan preview, working surfaces, cards, forms, Navigator dimensions, regulatory console, and footer with finer depth and hierarchy.
• Added subtle tonal variation, inner highlights, premium shadows, and restrained motion while preserving accessibility and reduced-motion behavior.
• Preserved the Version 23 information architecture, workflow logic, evidence controls, privacy model, and content depth.

Expanded Navigator readiness model

• Expanded Role, Milestone, Target date, Deliverable, Dependencies, Risk, Evidence, and Action plan into taller information tiles.
• Moved the former hover-only explanations directly into each tile so the readiness logic is visible without interaction.
• Added subtle hover/focus motion and a distinct Action Plan output treatment while retaining reduced-motion support.

Visitor-first purpose and readiness-gap messaging

• Rewrote the landing-page first impression around the specific problem CDRN solves: hidden decisions, dependencies, handoffs, risks, reviews, and evidence gaps that can still block a milestone.
• Made the intended cross-functional audience explicit in the hero and replaced generic proof bullets with Detect → Align → Prove outcome cues.
• Reframed the project-origin section around the distinction between progress and true readiness, and added a concise “readiness layer” explanation that differentiates CDRN from trackers, SOPs, standards, and approved study documents.
• Reframed the CDOR preview as the path from milestone context to an executable readiness plan.
• Added CDRN and Readiness gap definitions to the public glossary and corrected the homepage Field Guide count to 21.
• Updated homepage metadata and About-page purpose language to match the visitor-first positioning.

Content depth, source transparency, and cross-functional consistency

• Deepened flagship Field Guides and added Late Change Impact.
• Made guide handoffs/application steps topic-specific and added review/source metadata.
• Added regional FDA, EMA/EU, PMDA, Health Canada, MHRA, and TGA reference entry points; updated ICH E6(R3)/eCTD context.
• Added public methodology, citation, changelog, and glossary pages.
• Added adjustable Navigator checkpoint dates and configurable calendar reminders.
• Removed programming-role preselection from homepage scenarios and improved related-plan prefill.
• Clarified Practice Exchange provenance and regulatory-scenario cross-functional wording.
• Simplified feedback flow and exposed validation status without claiming independent review that has not occurred.
• Expanded the Field Guide to 21 topics, including external/vendor data, data cut/freeze, integrated submission/ISS-ISE, and CSR/Medical Writing handoffs.
• Added Safety / Pharmacovigilance and Submission Operations / Data Standards Navigator role lenses with explicit independent-review boundaries.
• Published the versioned CDOR Methodology v1.0 PDF and added a public professional-profile reference for creator provenance.

Landing-page composition correction

Restored the roomier readiness banner and v16-style example plan proportions, while keeping the current visual identity.

Evidence-ready architecture

Added first-party privacy-conscious usage-event hooks, structured practitioner feedback IDs, evidence-governance documentation, and private server storage design.

Cross-functional identity and Live Readiness Brief

Renamed the product Clinical Data Readiness Navigator, expanded the role model, and introduced an automatically updating situational brief before full plan generation.

Prototype to operational platform

Established Navigator, browser-local Workspace, Field Guide, Practice Exchange, Resources, Regulatory Scenarios, accessibility/privacy foundations, and the CDOR framework.