Skill UI

This skill transforms a refined research proposal or method idea into a detailed, structured, and claim-driven experiment roadmap. It helps users determine exactly which experiments to run, in what order, and what evidence is needed to support the core claims of a research paper, ensuring focus on novelty, simplicity, and genuine contribution.

This skill generates and ranks publishable research ideas based on a broad direction provided by the user. It systematically validates concepts through pilot experiments and model reviews, optimizing for GPU efficiency and novelty. Suitable for brainstorming sessions, academic exploration, and identifying workable research paths.

Orchestrate a complete idea discovery workflow that goes from a broad research direction to validated, pilot-tested ideas. Chains literature survey, idea generation, novelty checking, critical review, and research refinement into a single automated pipeline, producing a ranked idea report, a final proposal, and an experiment plan.

Orchestrates a robotics-specific literature survey, novel idea generation, novelty check, and critical review for embodied AI and robotics research. Helps turn a broad robotics direction (manipulation, locomotion, navigation, drones, humanoids) into benchmark-grounded, simulation-first pilot ideas. Use when you need robotics idea discovery or sim2real topic selection.

This skill verifies the novelty of research ideas against recent literature. It extracts core technical claims, searches academic databases and preprints, and cross-model verifies findings to determine if the idea is truly new or overlaps with existing work. It provides a structured report with a novelty verdict (Proceed, Proceed with Caution, Abandon).

Turns research ideas into complete grant proposals for KAKENHI, NSF, NSFC, ERC, DFG, SNSF, ARC, NWO, and generic formats. Supports novelty checking, structural design, drafting, simulated peer review, revision, budget planning, timeline development, and agency-specific formatting with multilingual output.

This system treats ideas like living species on an 'evolution island.' Instead of simple storage, it runs structured, multi-round iterations simulating biological processes: competition, crossbreeding, and mutation. Users input initial concepts, and the tool systematically evaluates them using a 6-dimensional scoring framework (Novelty, Feasibility, Value, etc.). It helps users deepen viability, discover unexpected connections between disparate ideas, and prioritize the most robust concepts for development and investment.

Evaluate patentability by analyzing novelty and non-obviousness against prior art. Supports US, CN, and EPO standards. Includes claim drafting, anticipation analysis, and cross-model verification to determine if an invention is patentable.

A complete patent drafting pipeline from invention description to jurisdiction-formatted filing documents. Supports CNIPA, USPTO, and EPO, covering prior art search, novelty checking, claims drafting, specification writing, examiner-style review, and jurisdiction-specific formatting for invention patents and utility models.

This skill simulates an external patent examiner to critically review patent applications. It analyzes claims and specifications for clarity, novelty, and non-obviousness. The process involves generating formal office actions, suggesting specific amendments, and iterating through multiple review rounds to improve patentability scores. It uses advanced reasoning models to provide rigorous feedback on patent claims and specification quality.

This tool facilitates comprehensive prior art searches across global patent databases (Google Patents, Espacenet) and academic literature (Google Scholar, arXiv). It guides users through extracting core inventive concepts, performing targeted searches using keywords and IPC/CPC codes, and analyzing the results to assess novelty, overlap risk, and general freedom-to-operate.

This advanced skill provides comprehensive patent prior-art and landscape intelligence. Instead of generic searches, it forces the user to commit to one of five specialized sub-use-cases: Novelty Search, Freedom-to-Operate (FTO), Competitive Landscape, Acquisition Diligence, or Litigation Prior-Art. It searches multiple global databases (Google Patents, Espacenet, USPTO) and outputs a detailed, actionable DOCX report with strategic recommendations, risk assessments, and technical verdicts.