The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses shell commands to drive specialized tools and research scripts. Specifically, it uses chrome-use to drive the user's local browser and python3 to execute internal scripts such as emerging_scan.py and auto_research.py. These operations are well-documented, localized to the skill's own environment, and serve the primary purpose of data gathering and skill optimization.
[EXTERNAL_DOWNLOADS]: The skill fetches financial data and news headlines from established public services. It uses yfinance to download price data and standard HTTP requests to fetch Google News RSS feeds for narrative tracking. It also accesses financial publications (WSJ, FT, Seeking Alpha) and SEC EDGAR. All external sources are well-known, authoritative financial data providers or standard technical registries (PyPI), fitting the skill's research-oriented purpose.
[SAFE]: The skill implements a robust 'Skeptic Filter' and explicitly forbids auto-trading. It uses a nomination-only model where findings are routed to a separate quorum for judgment, ensuring that it never makes final financial decisions autonomously. The 'Auto-Research' loop is a transparent, metrics-driven mechanism for refining the agent's reasoning capabilities based on a human-defined rubric.

trend-stock-research