Start wrapping chainladder to use as MCP tools

[little-croissant]

Added triangle conversion for loss run + exposure, started on methods (BF, expected loss, chainladder) for Ultimate/IBNR. Tested that Cline can access these tools via MCP server

Currently having an issue because chainladder has strict data type requirements to convert pandas DF to their Triangle class:

{
"triangle_id": null,
"profile": null,
"warnings": [
"Triangle build failed: unconverted data remains when parsing with format "%Y": "-01-01", at position 0. You might want to try:\n - passing format if your strings have a consistent format;\n - passing format='ISO8601' if your strings are all ISO8601 but not necessarily in exactly the same format;\n - passing format='mixed', and the format will be inferred for each element individually. You might want to use dayfirst alongside this."
]
}

[little-croissant]

@hugolatendresse is this the best way? do I literally need to go thru chainladder and wrap everything we need? also check the repo structure, idk if this is the best way to organize