Point your AI at this atlas

You are a research assistant. Fetch these URLs and use ONLY their JSON to answer:
https://tradefilings.com/api/v1/stock/TICKER
https://tradefilings.com/api/v1/stock/TICKER/ownership
https://tradefilings.com/api/v1/stock/TICKER/holders
Replace TICKER with the symbol I give you. Then write a 200-word institutional-ownership
brief: holder count and its change, largest holders by reported value, new buyers and exits,
and the momentum label. Rules: cite the reporting period next to every figure; reported
values are 13F values, not AUM; the data covers tracked 13F managers only; no advice, no
predictions. My ticker: 

Fetch these URLs and use ONLY their JSON:
https://tradefilings.com/api/v1/manager/SLUG
https://tradefilings.com/api/v1/manager/SLUG/holdings
https://tradefilings.com/api/v1/manager/SLUG/changes
(Find SLUG via https://tradefilings.com/api/v1/search?q=MANAGER+NAME first.)
Write a factual profile: reported 13F value with its period, position count, top-10
concentration, the five largest positions with portfolio weights, and this quarter's
notable new positions and exits. Never call the total "assets under management". A 13F
shows only reportable U.S. long positions — say so. No advice. Manager: 

Fetch: https://tradefilings.com/api/v1/screener?d_holders_min=5&new_min=5&limit=25
Using ONLY that JSON, list securities where tracked institutional ownership broadened this
quarter: ticker, holder count, holder change, new positions, momentum label. State the
reporting period from the meta. Note that this is a historical description of filings among
tracked managers, not a recommendation or prediction. Then suggest 2-3 refinement filters I
could add (sector, streak_min, holders_min) and show the modified URLs.

Fetch these URLs and use ONLY their JSON:
https://tradefilings.com/api/v1/insiders?ticker=TICKER
https://tradefilings.com/api/v1/stock/TICKER
List open-market insider purchases (code P only — never count awards or exercises as buys)
with dates and values, then set them beside the tracked institutional holder count and its
change. Present both as independent facts with their own dates; do not merge them into a
combined "signal" or a recommendation. Ticker: 

Why this shape: language models invent plausible ownership numbers when asked from memory. These prompts force the model to fetch our cached, attributed data first and reason only over it — the same rule our own Research Desk follows.