The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/lookup.mjs utility fetches documentation metadata from https://docs.thefintechbuilder.com/reference/payload.json and versioning information from https://docs.thefintechbuilder.com/version.json to provide reference functionality.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external financial market data (trades, bars, quotes), creating an attack surface for indirect prompt injection.
Ingestion points: Market data is ingested and adapted into library-specific shapes from various sources (REST APIs, CSV, websockets) as described in references/ingestion.md.
Boundary markers: The skill mandates the use of validation guards and explicit branching on validity results, as documented in SKILL.md and references/recipes.md.
Capability inventory: The library functions are zero-dependency mathematical utilities; the scripts/lookup.mjs script is limited to JSON fetching and caching in the temporary directory.
Sanitization: Input data is sanitized using dedicated validation modules including ohlc-consistency-validator, duplicate-trade-resolver, and hampel-bad-tick-filter as mapped in references/ingestion.md.
[DYNAMIC_EXECUTION]: The library exposes a runner function in its root export that allows algorithms to be dynamically loaded and executed by their catalog ID, enabling registry-driven workflows within the local package scope.

fintech-algorithms