Use this tool to answer questions about historical index membership — e.g. "Was Company X in the S&P 500 on date Y?" or "Which companies were in the Russell 2000 on 2010-01-01?" Use this INSTEAD OF `search_companies` when the question involves a specific historical date or asks whether a company was an index member at a point in the past. `search_companies` only returns current membership snapshots and cannot answer historical membership questions.
Returns a survivorship-free universe of companies valid on a specific as_of_date: only companies that existed and were index members on that exact date — no hindsight contamination. Supports SP500, RUSSELL1000, RUSSELL2000, RUSSELL3000 via index_membership.parquet (accurate join/leave dates with [) interval semantics). To check a single company's membership, pass its ticker and the target date; if the company appears in the response it was a member, if absent it was not.
Returns per company: CIK, ticker, name, sector, industry, SIC code, plus per-row membership confidence (high/medium/low). Check `_meta.pit_safe`: true only when every matched row is high-confidence; medium/low rows downgrade it to false — treat low-confidence rows with caution for backtest use.
NOTE: `sector` is SIC-derived (GICS-aligned labels via sic_to_sector.csv), not licensed GICS — industrial conglomerates may map differently. Treat as a screening bucket, not an authoritative GICS label.
Use as the first step of a quantitative backtest before calling `get_compute_ready_stream` to pull Parquet data for the universe. Returns empty array (with error detail) if the date is out of range or the index_membership data has no coverage for that date. Available on every plan — sample tier returns the subset covered by the sample bucket.
