Capture city + LinkedIn on card intake; sharpen the transcription prompt

The card transcription prompt now reads emails/URLs/phones character-by-character,
explicitly forbids autocompleting toward a plausible domain (the mara.com ->
marac.com failure), and emits labeled lines (which also feeds the field extractor
cleaner input).

The extractor gains city + linkedin_url. city is a plain field (low-harm if wrong;
the human sees it on the card). linkedin_url follows the email-integrity rule: kept
only if it literally appears in the source / a revise instruction, never minted -- a
wrong profile URL points at the wrong person. Both flow to the contact via the
existing log-communication upsert (city also syncs to the grid contact pill).

Phone is intentionally NOT included yet: the bot's write path can't store it until a
small server-side change lands (next s9pk). See the matrix-intake guide.

backend/matrix_intake/crm_client.py

@@ -162,6 +162,10 @@ def build_commit_payload(proposal):
         "name": proposal.get("contact_name") or proposal.get("investor_name") or "",
         "email": proposal.get("contact_email") or "",
         "title": proposal.get("contact_title") or "",
+        # city + linkedin_url are already honored by the server's contact upsert
+        # (_upsert_contact_from_fundraising); city also syncs to the grid contact pill.
+        "city": proposal.get("city") or "",
+        "linkedin_url": proposal.get("linkedin_url") or "",
     }
     note = proposal.get("note") or ""
     # The CRM's grid note line uses subject-or-body for its one-line summary, so a non-empty