No user-submitted related servers found.
Scored across 27 tools
The tools cluster into setup, ingestion, simulation, testing, and analysis, but several boundaries are fuzzy: market_test vs run_test vs study all orchestrate simulated tests, backtest vs holdout_backtest both score against real results, and ads_import vs actuals_import both ingest exports. Multiple entry points (understand, quickstart, research) also require reading descriptions to know which to use first. The descriptions do help, but an agent could easily misselect among the test and backtest tools.
All names are snake_case, which is consistent at the casing level. However, the semantic pattern is mixed: some are noun-only (ideas, results, pitch, study), some are verb-only (doctor, setup, calibrate, predict), and others follow verb_noun (run_test, list_runs, start_backend). It is readable but not a predictable convention.
At 27 tools, the surface is heavy for the apparent scope and includes many advanced or overlapping entry points (market_test, run_test, study, email_test, website_test, pitch, cli). While each has some role, the count exceeds the well-scoped 3-15 range and increases selection cost.
The lifecycle is broadly covered: setup/doctor, document understanding, research, ad/actuals ingestion, persona simulation, modular and all-in-one tests, email and website tests, backtests, calibration, prediction, dashboards, and raw CLI access. Minor gaps remain around managing or deleting runs/projects, editing personas, and exporting reports, but core workflows are well supported.