Skill UI

Emulates Liu Cixin's hard science fiction creative thinking, offering frameworks for macro-detail narrative, scientific concept-driven social conflict, technological optimism, and cosmic sociology. Ideal for generating sci-fi story ideas, worldbuilding, and dark forest-style plot conflicts.

This workflow mandates external verification for all non-trivial factual claims, ensuring accuracy in high-stakes domains. It is essential when researching current events, legal matters, scientific claims, or providing statistics. The process requires decomposing tasks, searching external sources, and providing immediate, inline citations to support every claim, overriding reliance on internal, potentially outdated knowledge.

This internal, deterministic gate enforces academic rigor before manuscript compilation. It audits the input, rejecting unsupported empirical claims, missing citation verdicts, or concrete outcome numbers that lack user evidence. It ensures that generated scientific papers only present validated results, blocking unsupported quantitative or qualitative assertions.

A comprehensive knowledge base covering the theoretical fundamentals and advanced algorithms of Deep Reinforcement Learning (Deep RL). It details core concepts like Bellman equations, policy gradients, and the trade-offs between policy optimization and Q-learning. The guide analyzes leading algorithms such as PPO, SAC, TRPO, and DDPG, and provides rigorous best practices for debugging and running scientifically sound RL experiments.

Retrieve, version, and publish scientific datasets with DataLad and git-annex, and capture reproducible computational provenance using datalad run, rerun, and containers-run for OpenNeuro, DANDI, and other DataLad datasets.

A discovery guide for scientific research and analysis skills, covering methods, data provenance, and evidence validation for reproducible results.

This skill implements a five-stage scientific thinking protocol for AI agents to prevent unverified guesses. It guides agents to formulate falsifiable hypotheses, seek evidence, and draw calibrated conclusions, especially for high-stakes or ambiguous questions. It includes modes for incidents, simple facts, estimates, and deep analysis, ensuring rigorous reasoning before answering.

Master modern Julia 1.10+ development with performance optimization, multiple dispatch, and production-ready practices. Covers scientific computing, data science, machine learning, web development, and package management. Focus on type stability, modern tooling, testing, and CI/CD pipelines for high-performance applications.

This skill generates scientifically accurate plots from NEXRAD Level II/III site scans and decoded radar mosaics. It ensures correct radar geometry, map grids, units, quality masks, timestamps, and data provenance. Suitable for meteorological analysis, it supports plotting reflectivity, radial velocity, dual-polarization moments, and various views like PPI, RHI, and cross-sections. It validates data integrity before visualization to prevent misleading interpretations.

Retrieve NOAA NEXRAD radar and GOES satellite products from public cloud archives, select scans by observation time, and validate downloaded scientific files.

Create publication-ready matplotlib figures for academic papers, slides, and reports. Covers bars, trends, scatter, heatmaps, multi-panel layouts, with house style, print/vector export conventions, and parity with figures4papers demos.

Search PubMed biomedical literature via NCBI E-utilities: build MeSH and field-tag queries, filter by publication type and date, retrieve PMIDs, abstracts, and citations for systematic reviews and literature monitoring.