What problem does it solve? Manually summarizing disclosed insider or institutional transactions across sources is error-prone, and it is easy to confuse holdings snapshots with actual trades or to silently copy trades without understanding direction, net flow, and disclosure lag. ## Core Features & Use Cases - Transaction Aggregation: Reads a local JSON array of disclosed transactions (date, side, source, value) and aggregates buys, sells, and net flow by asset and source via the AlphaGBM API. - Strict Input Validation: The bundled Python runner rejects malformed files, wrong field names (e.g. disclosedAt or action), non-positive amounts, and mixed assumptions before any charged request is made. - Safe Access Controls: Requires explicit --confirm-usage approval before consuming account allowance, keeps the API key in the environment only, and fails closed when the backend is unsupported. - Use Case: You collect public 13D/G or insider disclosure records for NVDA into a JSON file and ask the AI to summarize direction, net flow, sources, and limitations without inferring undisclosed positions. ## Quick Start Ask the AI to use AlphaGBM smart money tracking to summarize your disclosed transactions JSON file by asset and source, showing buys, sells, net flow, and limits, approving allowance use first.
