feat: expose filename on list_messages / get_message_context

[fenrith]

Feature request: expose filename on message tools

list_messages / get_message_context return media_type for document
attachments, but not the original filename. The bridge does store it
(messages.filename in messages.db, populated from Message.filename /
msg_to_dict in whatsapp.py), it's just not included in the dict returned
to MCP clients — msg_to_dict
only re-exposes filename when media_type == "reaction" (as
reaction_to_message_id).

Use case: filtering/identifying documents by type (e.g. "find PDFs sent
to this group") currently requires knowing the media is a generic
"document" and then reading messages.db directly to get the actual
filename/extension — which works, but means going around the MCP interface
entirely for something the data already has.

Suggestion: add filename to the msg_to_dict output for non-reaction
media types (it's already there in the Message dataclass, so this looks
like a small change), or alternatively expose it via a small addition to
download_media's response.

Happy to open a PR if a maintainer confirms this is a change you'd want —
wasn't sure if filename was deliberately withheld for privacy/size reasons
in the current design.