Second pass with the avoid-ai-writing detector

The skill's detector scored the README, PARITY.md and the extracted
comment prose. Fixed what it flagged: hollow intensifiers, "in order
to", "additionally", "the whole point", bold sentences in docs, and
bold that had wrapped across comment lines. Comment prose scores a
maximum of 4 out of 100 across the repo.

One sweep rule mangled four comments ("which is the entire point of X"
lost its clause); repaired, and the rule now keeps the clause.

Comments and strings only. 1208 tests pass, counted from the XML.

Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>

engine/src/test/kotlin/io/flamenet/e2e/LabelTest.kt

@@ -9,8 +9,8 @@ import kotlin.test.assertTrue
 /**
  * The security lattice, and the two Bell–LaPadula rules it exists to enforce.
  *
- * **The encoding vectors were computed independently, not read off this
- * implementation.** The ten bytes are authenticated data for every construction
+ * The encoding vectors were computed independently, not read off this
+ * implementation. The ten bytes are authenticated data for every construction
  * that uses a label, so an encoding that differs between platforms does not warn
  * it produces a decryption failure, or plaintext that quietly belongs to the
  * wrong sensitivity. A vector taken from the code under test would pin whatever
@@ -148,8 +148,8 @@ class LabelTest {
     // ---- Writing ---------------------------------------------------------
 
     /**
-     * **Writing is gated by the same range check as reading, and the
-     * -property is NOT enforced, on either platform.**
+     * Writing is gated by the same range check as reading, and the
+     * -property is NOT enforced, on either platform.
      *
      * Bell–LaPadula says somebody cleared to read sensitive material must not
      * write *below* their clearance, because they would carry what they read