feat(learning): chair-feedback style corrections auto-flow to the writer (P1 #6)

The chair-feedback chain was DEAD: 27 feedback rows captured, 0 ever became a
lesson the writer reads (no weekly-analysis job ever existed). The chair's own
corrections — the most authoritative signal of all, and exactly "learn from every
returned draft" — went nowhere.

decision_lessons can't carry them: it's FK-coupled to a style_corpus row (a signed
final), which a case-in-progress doesn't have. So chair STYLE feedback instead
rides the discussion_rules channel that already reaches the writer for all blocks —
the same path /training promote uses.

- db.append_global_rule: the locked read-modify-write append, extracted from web
  `_append_methodology_override` into one shared impl (G2). The web function is now
  a thin wrapper that seeds defaults; chair-feedback calls it directly.
- record_chair_feedback (MCP tool): a STYLE-category correction (style/wrong_tone/
  wrong_structure) with a lesson_extracted flows immediately to discussion_rules.
  The chair IS the gate — no separate approval (INV-LRN1 graduated gate). SUBSTANCE
  feedback (missing_content/factual_error/other) is case-specific → recorded only.

Invariants: INV-LRN1 (chair-authored style = highest authority, flows; substance
not auto-flowed), G2 (single append impl shared by promote + feedback), INV-LRN5
(style channel only).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

mcp-server/src/legal_mcp/services/db.py

@@ -2875,6 +2875,48 @@ async def get_style_patterns(pattern_type: str | None = None) -> list[dict]:
     return [dict(r) for r in rows]
 
 
+async def append_global_rule(
+    category: str, key: str, items: list[str], seed_if_missing: list | None = None,
+) -> int:
+    """Append items to a _global appeal_type_rules list value (the writer-consumed
+    methodology channel), idempotently and under a row lock. Returns how many NEW
+    items were added (existing duplicates are skipped). Single locked read-modify-
+    write (MET-2/3) so concurrent appends can't drop items. Shared by the /training
+    promote gate (web `_append_methodology_override`) and chair-feedback auto-flow
+    (G2 — one append implementation)."""
+    pool = await get_pool()
+    async with pool.acquire() as conn:
+        async with conn.transaction():
+            row = await conn.fetchrow(
+                "SELECT rule_value FROM appeal_type_rules "
+                "WHERE appeal_type = '_global' AND rule_category = $1 AND rule_key = $2 "
+                "FOR UPDATE",
+                category, key,
+            )
+            if row:
+                current = row["rule_value"]
+                if isinstance(current, str):
+                    try:
+                        current = json.loads(current)
+                    except (json.JSONDecodeError, TypeError):
+                        current = []
+            else:
+                current = list(seed_if_missing or [])
+            if not isinstance(current, list):
+                current = []
+            added = [s for s in items if s and s not in current]
+            if not added and row:
+                return 0
+            merged = current + added
+            await conn.execute(
+                "INSERT INTO appeal_type_rules (id, appeal_type, rule_category, rule_key, rule_value) "
+                "VALUES (gen_random_uuid(), '_global', $1, $2, $3::text::jsonb) "
+                "ON CONFLICT (appeal_type, rule_category, rule_key) DO UPDATE SET rule_value = $3::text::jsonb",
+                category, key, json.dumps(merged, ensure_ascii=False),
+            )
+            return len(added)
+
+
 async def get_methodology_overrides(category: str) -> dict:
     """Chair's /methodology edits for one category (golden_ratios / discussion_rules /
     content_checklists). Returns {rule_key: parsed_value}. These OVERRIDE the hardcoded