"author:Luissalet" matching MCP servers:

Enables language models and users to perform exact arithmetic, symbolic math, unit/date conversions, statistics, and read-only SQL queries over registered files, with every operation logged by a citable ID that humans can re-run.

Enables assistants to reach a local market-research lab over MCP, including provenance-tagged market snapshots, falsifiable theses with invalidation rules, strategy backtesting with overfitting checks, committee reviews, portfolio analysis, and report export.

Enables a local assistant to drive a full data workbench: ingest messy files, clean them with versioned and undoable steps, enforce quality rules, build charts and dashboards as code, train quick models, and ask questions of the data—all on the user's own machine.