init project

contracts/phase9/reference/FINDINGS_REPORT_v1.0.md

@@ -0,0 +1,37 @@
+# Findings Report
+
+**Version:** 1.0  
+**Evidence status:** Pre-test expert review and automated validation only  
+**Participant sessions completed:** 0
+
+## Executive assessment
+
+The Phase 8 prototype is ready for pilot testing. The Phase 7 baseline exposed several behavioral gaps that would have invalidated primary-flow testing: inert form submission, no product-tour interaction behind a “Watch product tour” promise, incomplete mobile drawer focus containment, and no durable theme preference. These were corrected in the Phase 8 prototype.
+
+No claim is made that users understand the product, find the CTA, or complete the form. Those are hypotheses until representative participants attempt the tasks.
+
+## Pre-test findings
+
+| ID | Evidence | Finding | Severity | Status |
+|---|---|---|---|---|
+| P8-001 | Expert + code review | Demo submission in Phase 7 prevented default and produced no success, failure, or recovery state. | S0/C | Fixed in prototype |
+| P8-002 | Expert + code review | “Watch product tour” navigated to a workflow section rather than opening a tour, creating an expectation mismatch. | S1/C | Fixed with four-step tour |
+| P8-003 | Expert + code review | Mobile drawer moved focus initially but did not contain focus or make background content inert. | S0/C | Fixed in prototype |
+| P8-004 | Expert + code review | Theme existed in URL state but did not persist reliably when revisiting without the query parameter. | S1/C | Fixed with local persistence |
+| P8-005 | Expert review | Multi-location support is implied by branch/location copy but is not an approved capability claim. Users may over-infer. | S2/C | Test explicitly; content approval pending |
+| P8-006 | Expert review | Final legal consent, alternate failure contact, CRM routing, and spam controls are unresolved. | S1/C for production | Prototype limitation; open |
+| P8-007 | Automated browser test | EN/FR/AR direction, tour completion, form validation/success, locale hash preservation, theme persistence, and mobile focus return pass the scripted browser checks. | S4/A | Passed |
+
+## Hypotheses requiring participant evidence
+
+1. The first viewport communicates rental operations and reservation-to-return scope within five seconds.
+2. Orange CTA hierarchy yields unprompted demo discovery.
+3. The four-step tour improves workflow comprehension without being mistaken for a full product demo.
+4. Arabic users understand menu origin, control order, and mixed LTR identifiers.
+5. Users distinguish “pricing factors” from public pricing.
+6. Users do not mistake neutral integration and security evaluation content for confirmed capability or certification.
+7. Form error recovery preserves confidence and entered data.
+
+## Participant-results section
+
+After sessions, add cohort counts, task-level completion, median time, error and rescue rates, abandonment, confidence, accessibility outcomes, repeated behavior patterns, anonymized quotes where consented, issue table, and recommended revisions. Report locale results separately before any aggregate.

contracts/phase9/reference/PHASE_8_HANDOFF_v1.0.md

@@ -0,0 +1,19 @@
+# Phase 8 Handoff — Interactive Prototype and Usability Testing
+
+Phase 8 is test-ready.
+
+Use `08_INTERACTIVE_PROTOTYPE_USABILITY_TESTING_v1.0.md` as the authoritative research specification, `prototype/` as the clickable test build, and `DECISION_LOG_v8.0.md` as the cumulative decision record.
+
+The prototype provides:
+
+- EN/FR/AR content with true Arabic RTL;
+- Light/Dark/System selection with persistence;
+- context-preserving locale changes;
+- demo form validation, simulated service failure, confirmation, and reset;
+- four-step product tour;
+- mobile navigation with focus containment, inert background, Escape handling, and focus return;
+- session-local, non-PII research events and export in `?test=1` mode.
+
+Research must begin with three pilot sessions, then the balanced core and accessibility cohorts defined in `PARTICIPANT_PROFILE_v1.0.md`. Do not mark Phase 8 accepted until the thresholds in `SUCCESS_METRICS_v1.0.csv` pass and every Critical/High issue is fixed and retested.
+
+Unresolved product, recruitment, market, analytics, CRM, legal, and content matters remain in `OPEN_QUESTIONS_v1.0.md`.

contracts/phase9/reference/PRIORITIZED_REVISIONS_v1.0.csv

@@ -0,0 +1,8 @@
+Revision_ID,Finding_ID,Revision,Severity,Status,Owner,Target,Retest
+REV-001,P8-001,"Implement validation, failure, success, and reset states",S0,Done,Prototype,Phase 8 v1.0,Automated + participant retest
+REV-002,P8-002,Replace misleading tour anchor with four-step tour dialog,S1,Done,Prototype,Phase 8 v1.0,Tour task
+REV-003,P8-003,"Add drawer focus containment, inert background, and focus return",S0,Done,Prototype,Phase 8 v1.0,Keyboard + screen reader
+REV-004,P8-004,Persist theme and restore it after reload,S1,Done,Prototype,Phase 8 v1.0,Theme task
+REV-005,P8-005,Test and then clarify multi-location capability language using approved product evidence,S2,Pending evidence,Product + Content,Before beta,Affected cohort retest
+REV-006,P8-006,"Approve consent, CRM routing, spam protection, alternate contact, retry and duplicate handling",S1,Open,Legal + Sales Ops + Engineering,Before production,Integration test
+REV-007,Analytics open question,Map provisional events to approved analytics/consent platform,S2,Open,Analytics + Legal,Before beta,Event QA

contracts/phase9/reference/SUCCESS_METRICS_v1.0.csv

@@ -0,0 +1,13 @@
+ID,Metric,Definition,Threshold,Collection,Segmentation
+M-01,Five-second comprehension,Correct category + audience + reservation-to-return scope,>=80% overall; no locale <70%,Exact unprompted response,Per locale
+M-02,Primary CTA discovery,Book demo activated without prompt,>=90% within 10 seconds,Time and first action,Per locale/device
+M-03,Product tour completion,All four steps completed and sequence paraphrased,>=85%,"Completion, time, errors",Per locale
+M-04,Module discovery,Vehicle inspections located and explained,>=85% without rescue,"Completion, path, time",Per role
+M-05,Multi-location interpretation,Finds evidence and separates confirmed from unresolved,>=80%; overclaim <=10%,Response and cited evidence,Multi-location vs single
+M-06,Pricing discovery,Finds quote model and >=2 factors,>=85%,Completion and time,Per locale
+M-07,Locale context,Switch preserves current section and task,100% technical; >=90% participant continuation,"Hash, section, errors",EN↔FR; EN↔AR
+M-08,Theme persistence,Selection survives reload,100%,State before/after reload,Browser/OS
+M-09,Demo completion,Reaches confirmation,>=90%,Completion and abandonment,Per locale/device
+M-10,Validation recovery,Corrects error without losing other data,>=85%,"Errors, recovery time, rescue",Per locale/access method
+M-11,Keyboard critical path,Completes navigation/theme/demo with keyboard,100%; zero traps,Focus order and completion,Desktop/mobile keyboard
+M-12,Confidence,Post-task rating,Median >=4/5; no locale <3.5,Rating plus reason,Per locale

contracts/phase9/reference/VALIDATION_RESULTS_v1.0.md

@@ -0,0 +1,35 @@
+# homePageCar — Phase 8 Validation Results
+
+**Version:** 1.0  
+**Package status:** PASSED for prototype readiness  
+**Usability acceptance status:** NOT YET TESTED WITH PARTICIPANTS
+
+## Automated package validation
+
+- 63 static package and behavior-contract checks passed.
+- Required deliverables, locale copy parity, form state copy, tour state copy, interaction hooks, privacy constraints, and decision continuity are present.
+- JavaScript syntax and browser execution completed successfully in Chromium.
+
+## Scripted browser validation
+
+Six checks passed:
+
+1. English homepage and four-step product-tour completion.
+2. French demo-form validation recovery and confirmation.
+3. Locale switch preserves the pricing-section context.
+4. Explicit theme selection persists after reload/revisit.
+5. Arabic dark mobile drawer opens in RTL, moves focus, closes with Escape, and returns focus.
+6. Simulated service failure retains entered values and presents a persistent error.
+
+The execution environment blocks browser navigation to local HTTP and file URLs by administrator policy. The browser harness therefore injects the exact packaged HTML, CSS, copy, and JavaScript into Chromium and supplies deterministic query state in memory. This validates behavior but does not replace manual testing from a normal local server, real browser storage, assistive technology, zoom, virtual keyboards, or network conditions.
+
+## Manual gates still required
+
+- Moderated pilot and core participant sessions.
+- Native Arabic and French linguistic review.
+- NVDA, VoiceOver, TalkBack, and real keyboard/switch-user testing.
+- 200% zoom, 400% reflow, forced colors, reduced motion, and mobile virtual keyboard.
+- Real CRM submission, consent, duplicate handling, retry, spam protection, and analytics consent.
+- Final multi-location, pricing, integrations, security, legal, and evidence approvals.
+
+Automated checks establish that the prototype behaves as specified. They do not establish that people understand it, trust it, or refrain from inventing capabilities between the lines.