@roamer-remote

a-stocks-analysis maps natural language queries to structured stock market data retrieval across chinese a-shares and hong kong stocks, supporting price action, fundamentals, indices, and corporate communications with mandatory api key auth and explicit non-trading disclaimers.

debugging-reinforcement-learning covers seed management, trajectory replay, behavior analysis, and reward decomposition to isolate non-determinism sources and diagnose agent bugs in stochastic rl environments.

alcor-capability-evolver is a self-modifying agent skill that analyzes runtime history and applies constrained protocol-based improvements via git and npm. execution modes exist (review, mad dog loop, automated) but core procedures lack step-by-step clarity on what evolution actually does.

ocr-python wraps paddleocr for chinese and english text extraction from images and pdfs, with basic code examples but minimal guidance on error handling or production use.

Measure information entropy and redundancy in agent group communications. Use when user asks about agent communication efficiency, information redundancy, en...