docs: add ADR-007 (analytics tooling) and ADR-008 (playbook split rule) (#181)

Two decisions from the 2026-04-25 audit session that were recorded
in the dev journal but not as formal ADRs.

Co-authored-by: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Author: braboj

docs/decisions/007-analytics-tooling.md

@@ -0,0 +1,30 @@
+# ADR-007: Analytics tooling — GSC for tutorials, Plausible for Wuseria
+
+**Status:** Accepted
+**Date:** 2026-04-25
+
+## Context
+The tutorial site needed analytics to measure search performance and reader
+engagement. The options ranged from free (Google Search Console, Google
+Analytics) to privacy-focused paid tools (Plausible, Fathom).
+
+## Decision
+Use Google Search Console (GSC) for all Code with Branko tutorials. Reserve
+Plausible for commercial products (Wuseria) where privacy-first analytics
+justify the cost.
+
+## Alternatives considered
+- **Google Analytics**: Free and full-featured, but heavy JavaScript bundle,
+  cookie banners required, and privacy concerns for a developer audience
+- **Plausible for everything**: Clean and privacy-first, but costs money per
+  site — unnecessary for free tutorials where GSC covers the key metrics
+  (impressions, clicks, indexing)
+- **No analytics**: Leaves content decisions ungrounded — no way to know
+  which chapters attract readers or which queries drive traffic
+
+## Consequences
+- GSC is free and sufficient for tutorials — covers search performance,
+  indexing, and sitemap validation
+- No client-side analytics script on tutorial pages — faster load, no
+  cookie banner
+- Plausible cost is only incurred for commercial products

docs/decisions/008-playbook-split-rule.md

@@ -0,0 +1,32 @@
+# ADR-008: Playbook split rule — standalone pages for substantial recipes
+
+**Status:** Accepted
+**Date:** 2026-04-25
+
+## Context
+The playbook chapter was split from a single file into individual recipe
+pages for SEO (#129, #153). The question was which recipes justify their
+own page versus staying grouped in a parent file.
+
+## Decision
+A recipe gets its own page only if it has enough substance to stand alone —
+at minimum a command, an explanation, and a gotcha or variation. One-liner
+recipes stay grouped in their parent file.
+
+## Alternatives considered
+- **Every recipe gets its own page**: Maximises long-tail SEO surface, but
+  creates thin pages that dilute quality signals and fragment navigation
+- **Keep everything grouped**: Simpler structure, but long pages rank poorly
+  for focused queries like "how to remove a git submodule"
+- **Split by search volume**: Data-driven but requires keyword research
+  tooling and ongoing maintenance — too heavy for a one-person operation
+
+## Consequences
+- Recipes with multi-step workflows, scripts, or common gotchas become
+  standalone pages (e.g. git bisect, pre-commit hook, SSH setup, remove
+  submodule)
+- Simple one-liner recipes stay grouped in their parent topic file
+- The rule is content-driven, not data-driven — no dependency on keyword
+  tools
+- Issue #154 (improve thin recipes) should run before #153 (split into
+  pages) to ensure substance exists before splitting