Skill UI

Handles user prompts to send local files or images by emitting the required Weixin protocol block so the app can deliver them directly; report success only when the block and target file exist to avoid false claims.

Three-layer verification pipeline that extracts verifiable claims from AI output, searches the web for supporting or contradicting sources, flags hallucinations, and builds structured reports for human review.

A persistent, structured knowledge base designed to accumulate and connect all research artifacts—including papers, ideas, experiments, and claims—across the entire research lifecycle. Inspired by LLM Wiki patterns, it builds a comprehensive, evolving graph of knowledge by tracking typed relationships (e.g., supports, contradicts, extends), ensuring that findings are compounded and never re-derived. Ideal for complex, multi-stage scientific projects.

This guide outlines a mandatory, evidence-based protocol for declaring any work complete, fixing a bug, or submitting a pull request. It strictly mandates that all claims of success must be supported by fresh, verifiable evidence, such as passing test outputs, successful build logs, or comprehensive checklist verification. It eliminates reliance on assumptions, gut feelings, or trusting proxy reports.

Utilize advanced media intelligence to analyze audio, image, video, and text for synthetic manipulation and AI-generated content. This skill detects deepfakes, traces the original AI synthesis source, applies and detects watermarks, verifies speaker identity, and provides comprehensive media provenance analysis, ensuring media authenticity claims are rigorously supported.

This skill provides a systematic framework for evaluating the rigor of scientific claims, research methodology, and evidence quality. It guides users through assessing experimental design validity, identifying various types of biases (e.g., selection, cognitive), and applying established grading frameworks like GRADE or Cochrane Risk of Bias. Ideal for systematic reviews, peer review, and critical analysis of scientific literature.

This skill rigorously verifies all claims of completion, whether for a task, a bug fix, or a feature rollout. It prevents false assurances by demanding fresh, multi-faceted evidence, including full test suite results, comprehensive coverage reports, integration status, and validation commands. Use it before declaring a component ready for GO or accepting a sub-agent's success report.

This core skill drafts comprehensive and legally sound patent claims based on invention disclosures and prior art. It adheres to complex jurisdictional standards (US, EP, CN) and requires defining the precise legal scope of protection for both method and apparatus claims. Essential for the intellectual property pipeline.

This powerful skill audits the integrity and reliability of AI experimental results. It employs a cross-model review process to detect common forms of methodological fraud, including fake ground truth generation, suspicious score normalization, and insufficient scope reporting. Use this tool whenever you need an adversarial check on experimental claims to ensure the reported outcomes are robust and trustworthy.

This skill automates the complex process of creating formal drawing descriptions and reference numeral indexes for patent applications. It analyzes user-provided technical diagrams (PNG, JPG, SVG, PDF), identifies components, assigns sequential reference numerals according to standard patent rules (e.g., 100-199 for FIG. 1), and generates structured descriptions in both Chinese (CN) and English (US/EP). It also cross-references the generated index against patent claims to ensure completeness.

Automates the compilation of a complete patent application from source files into legally compliant, filing-ready formats for major global patent offices (CNIPA, USPTO, EPO). This tool handles the complex, jurisdiction-specific rules for claims structuring, technical field description, specification ordering, and abstract formatting, ensuring consistency across multiple national standards. Users can specify the required jurisdictions (China, US, Europe) and output format (Markdown or DOCX).

This bundle generates comprehensive diagnostic logs for troubleshooting OneNote Graph API issues. It captures critical data points—such as the `request-id`, `x-ms-ags-diagnostic` headers, full error bodies, and decoded JWT token claims—for both successful and failed API calls. Use it for self-diagnosis, verifying permissions, or preparing detailed support tickets for Microsoft support.