Fix docs MDX parsing and add PR build check

.github/workflows/docs-check.yml

@@ -0,0 +1,31 @@
+name: Docs Check
+
+on:
+  pull_request:
+    paths:
+      - 'docs/**'
+      - '.github/workflows/docs-check.yml'
+
+jobs:
+  build:
+    name: Build Documentation
+    runs-on: ubuntu-latest
+    defaults:
+      run:
+        working-directory: docs
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v6
+
+      - name: Set up Node.js
+        uses: actions/setup-node@v6
+        with:
+          node-version: 20
+          cache: npm
+          cache-dependency-path: docs/package-lock.json
+
+      - name: Install dependencies
+        run: npm ci
+
+      - name: Build site
+        run: npm run build

docs/docs/advanced/mutation-testing.md

@@ -143,6 +143,6 @@ A mutation is considered "detected" if **any** metric on **any** test input prod
 
 As a rule of thumb:
 
-- **>80% detection rate** -- strong evaluation coverage
+- **80%+ detection rate** -- strong evaluation coverage
 - **50--80% detection rate** -- acceptable but review undetected mutations
-- **<50% detection rate** -- evaluation metrics need improvement
+- **Under 50% detection rate** -- evaluation metrics need improvement

docs/docs/advanced/red-teaming.md

@@ -77,7 +77,7 @@ void agentShouldHandleBoundaryInputs(AgentTestCase testCase) {
 
 Built-in boundary tests:
 - Empty string
-- Extremely long input (>100k characters)
+- Extremely long input (100k+ characters)
 - Special characters and unicode
 - SQL/HTML/JSON injection strings
 - Null bytes and control characters

docs/docs/advanced/statistical-analysis.md

@@ -102,10 +102,10 @@ Stability assessments are based on the coefficient of variation (CV):
 
 | CV Range | Assessment |
 |---|---|
-| CV <= 0.05 | Highly stable |
-| CV <= threshold | Stable |
-| CV <= threshold x 2 | Moderately unstable |
-| CV > threshold x 2 | Highly unstable |
+| CV ≤ 0.05 | Highly stable |
+| CV ≤ threshold | Stable |
+| CV ≤ threshold x 2 | Moderately unstable |
+| CV &gt; threshold x 2 | Highly unstable |
 
 A warning is emitted if fewer than 3 runs are provided.