FABRIC-3.md §H.10: close pin-authority choke point -- session owns both directions

Decided: session_set_pinned()/session_is_pinned() (or equivalent) are the
sole read AND write path for pin state -- nothing, including existing
Stadium code, touches STADIUM_FLAG_PIN on the patron header directly
anymore. Not just a write-side guard.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

FABRIC-3.md

@@ -3749,10 +3749,13 @@ child each get their own separate dictionary — `DictEntry` ACL fields are alre
 scoped per-session. H.3's stack-of-cards model and H.5's word card/elevation trigger stand as
 decided, unchanged by this check.
 
-Two smaller, lower-risk items flagged in the same pass, not yet acted on:
-- **Pin-authority sync risk (H.2).** Session.pinned is authoritative over Stadium's
-  `STADIUM_FLAG_PIN` bit — needs a single choke-point API so nothing ever writes the Stadium
-  flag directly and lets the two drift out of sync. Not yet designed.
+Two smaller, lower-risk items flagged in the same pass:
+- **CLOSED 2026-09-02 — pin-authority choke point (H.2).** Session.pinned is authoritative
+  over Stadium's `STADIUM_FLAG_PIN` bit. Decided: **full choke point at the session level,
+  both directions** — both writing and reading pin state go exclusively through session-owned
+  functions (e.g. `session_set_pinned()` / `session_is_pinned()`); nothing, including existing
+  Stadium code, reads `STADIUM_FLAG_PIN` directly off the patron header anymore. Session is
+  the sole authority for both write and read, not just the write path.
 - **Elevation trigger not yet usable (H.5/H.7).** Rides Hermes messaging, which per this
   document's own standing completion criterion (D.1) isn't the real implementation yet.
   Correct on paper, not usable until that substrate exists — expected, not a flaw.