docs: recover Career Workspace research + strategy from feature/career-workspace

Bring the four Career Workspace documents onto main as the target architecture
for Phases 2-4, and point MASTER_IMPLEMENTATION_GUIDE.md at them. Taken from the
branch tip (later commits refined them). Pure additions — none previously existed
on main.

- cv-builder-competitor-deep-research.md (Novoresume, Reactive Resume, FlowCV,
  Teal, Enhancv, Canva, Resume.io, Kickresume; matrix; pricing intelligence).
- cv-builder-product-teardown.md
- career-workspace-product-strategy.md
- career-workspace-implementation-roadmap.md (F0-F5)

MASTER_IMPLEMENTATION_GUIDE.md v1.1: adds a Source-Of-Truth Documents section and
restates the "profile is the source of truth; documents reference snapshots" rule.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

docs/MASTER_IMPLEMENTATION_GUIDE.md

@@ -1,18 +1,21 @@
 # Jobjakt Master Implementation Guide
 
-Version: 1.0
+Version: 1.1
 
 ---
 
 # Purpose Of This Document
 
-This document is the primary source of truth for understanding and improving Jobjakt.
+This document is the primary source of truth for the product *direction*. For the
+*current system* and the *recovered Career Workspace architecture*, read the documents
+linked below — they are verified against the code and take precedence over any stub
+elsewhere in `docs/`.
 
 Before making significant changes:
 
-1. Read this document.
-2. Read the relevant linked documentation.
-3. Understand existing architecture.
+1. Read this document (direction and non-negotiables).
+2. Read the relevant linked documentation (below).
+3. Understand existing architecture — verify against the actual codebase.
 4. Make incremental improvements.
 
 Do not redesign based on assumptions.
@@ -21,6 +24,46 @@ Always verify against the actual codebase.
 
 ---
 
+# Source-Of-Truth Documents
+
+Read these before touching the areas they cover. Where one disagrees with the code,
+the code wins and the document is a bug to fix.
+
+## Current system (verified)
+
+- `docs/architecture/current.md` — the system as it actually is: projects, controllers,
+  data model, auth, AI pipeline, CI/CD, security posture. **Start here.**
+- `docs/application-discovery-report.md` — full audit and feature inventory.
+- `docs/phase-0-foundation-report.md` — what Phase 0 changed (Job entity, expanded
+  pipeline, AI-service lockdown, DateApplied history) and how it was verified.
+- `docs/implementation-roadmap.md` — the phased plan and settled product decisions.
+
+## Decisions
+
+- `docs/decisions/ADR-002-job-application-model.md` — Job (opportunity) separated from
+  JobApplication (pursuit); the reference-not-ownership rule.
+
+## Career Workspace — the recovered architecture (governs Phase 2–4)
+
+These four documents were recovered from the `feature/career-workspace` branch and are
+the **target architecture** for the Career Workspace. They, plus ADR-002, define how
+career data is stored and how generated documents reference it.
+
+- `docs/career-workspace-product-strategy.md` — vision, positioning, personas.
+- `docs/career-workspace-implementation-roadmap.md` — the F0–F5 execution plan.
+- `docs/cv-builder-competitor-deep-research.md` — Novoresume, Reactive Resume, FlowCV,
+  Teal, Enhancv, Canva, Resume.io, Kickresume; feature matrix; pricing intelligence.
+- `docs/cv-builder-product-teardown.md` — product teardown.
+- `docs/career-workspace-branch-assessment.md` — per-table recover/migrate/replace verdict
+  for the recovered Career/CV tables.
+
+**Non-negotiable, restated from those documents:** the master career profile is the single
+source of truth. Career data is never duplicated into CVs, applications, or variants — a
+generated document *references* a snapshot; it does not own the data. Job tracking remains
+the primary product; the Career Workspace is a bounded supporting domain.
+
+---
+
 # Product Identity
 
 ## Application Name