Skill UI

A comprehensive framework for designing, auditing, and reporting empirical results in recommender systems research. It covers best practices for handling temporal splits, mitigating evaluation bias (e.g., exposure/popularity), utilizing advanced off-policy estimators (IPS, SNIPS), and structuring the critical link between offline metrics and real-world A/B test results. Essential for academic submissions.

This guide is designed for systematic reviews and meta-analyses, providing a rigorous framework to stress-test the robustness and completeness of reported findings. It mandates advanced checks such as saturation evidence, risk-of-bias appraisal, and statistical diagnostics (e.g., heterogeneity, publication bias) to ensure conclusions are not artifacts of omission or methodological weakness. Essential for high-level academic research.

This skill bridges theoretical econometric designs (DiD, IV, RDD, SCM, DML) to concrete, audited statistical estimates. It orchestrates the necessary tool calls to fit models, generate comprehensive diagnostics, and report reliable findings, ensuring that every academic claim is backed by executable, traceable code. Essential for advanced causal inference research.

This skill provides comprehensive guidelines for conducting and reporting high-stakes quantitative data analysis in the social sciences. It emphasizes methodological rigor, requiring honest reporting of uncertainty, probing robustness (specs that could break the result), and ensuring full reproducibility. It covers advanced topics like design-based inference, cluster corrections, and handling missing data, ensuring results withstand expert review.

This comprehensive guide details the rigorous methodology for designing, executing, and reporting experimental results in Information Retrieval (IR). It covers best practices for selecting appropriate test collections, choosing defensible metrics, performing proper statistical significance testing (e.g., paired t-tests), ensuring baseline fairness, and proactively addressing modern pitfalls like LLM contamination. Essential for publishing high-quality, auditable research in IR/NLP.

A comprehensive guide for designing rigorous Monte Carlo simulation studies, particularly for methods-heavy academic submissions (e.g., SMR). This guide covers specifying the Data Generating Process (DGP) space, selecting non-negotiable competitor methods, choosing appropriate performance metrics (e.g., power, bias, coverage), and critically reporting the limitations and failure regimes of the proposed method to ensure maximum credibility and reproducibility.

This guide provides a comprehensive framework for designing rigorous and defensible evaluation sections for machine learning and information retrieval papers, particularly for top-tier conferences like WSDM. It covers the four critical evidence quadrants—Effectiveness, Validity, Efficiency, and Robustness—and details necessary protocols, such as temporal splits, proper baseline selection (including heuristics), ablation studies, and hardware-aware performance reporting. Use this when auditing or writing the experimental section of a research paper.

This utility is designed to automatically generate professional, formatted reports from structured data sources. Users simply provide the designated data input, and the system handles the entire report generation process, ensuring high consistency and quality across all resulting documents. It is ideal for data analysts and business intelligence users who require standardized reporting.

This skill audits a complete security or forensics case package to ensure its readiness for formal report handoff. It verifies scope metadata, evidence record integrity, traceability between Findings, Paths, and Evidence, and checks artifact hash consistency. It is a non-destructive, read-only review process designed to validate the defensibility and completeness of the investigation before documentation.

This skill analyzes local Q&A session JSONL files to produce a complete, auditable weekly report in Markdown format. It supports generating routine weekly summaries, comparing performance across consecutive time windows, and identifying key trends and risks. The output is a single, self-contained Markdown file designed to provide sufficient aggregates and structure for downstream documentation rendering (e.g., Feishu or whiteboards), without needing external data sources or APIs.

A comprehensive toolkit for the entire lifecycle of spreadsheet data, including creation, sophisticated editing, deep analysis, and visual reporting. It supports various formats (.xlsx, .csv) and is designed for professional tasks such as building complex financial models, generating reports with precise formatting, and analyzing raw tabular data while meticulously preserving formulas and structure.

This skill processes raw, unstructured materials (e.g., conversations, reports, controversies) into highly structured, reusable casebook entries. It focuses on identifying core facts, defining mechanism chains, and mapping responsibility links while ensuring rigorous data anonymization. It is designed for deep knowledge management and pattern recognition, helping users build a reliable, reusable knowledge base for future professional analysis, rather than generating immediate advice.