Use when main results pass result-to-claim (claim_supported=yes or partial) and ablation studies are needed for paper submission. Codex designs ablations from a reviewer's perspective, CC reviews feasibility and implements.
Analyze ML experiment results, compute statistics, generate comparison tables and insights. Use when user says "analyze results", "compare", or needs to interpret experimental data.
Draft patent claims for an invention. Use when user says "撰写权利要求", "draft claims", "写权利要求书", "claim drafting", or wants to create patent claims. The core skill of the patent pipeline.
Use Codex locally to update skills/index.json for Personal Skills MCP by reading only new or missing SKILL.md entries while preserving existing index entries.
深度学术研究与事实核查。Use when user says "深入研究", "深入核查", "深入核实", "深入检索", "深度调研", "深度研究", "deep research", "deep dive", "深入了解", "彻底调查", "全面核查", "深挖", or wants thorough multi-source literature investigation with PDF deep reading, cross-model adversarial review, and structured evidence audit. 注意：如果用户只是简单查文献/找论文/查参数，没有说"深入"或"深度"，请改用 quick-lit skill skill。本 skill 是重量级流程：文献检索 → PDF精读 → 交叉验证 → 对抗审阅 → 结构化报告。
Write detailed embodiment descriptions for patent specifications. Use when user says "撰写实施例", "write embodiment", "实施例描述", "detailed description", or wants to describe how to practice an invention.
Audit experiment integrity before claiming results. Uses cross-model review (GPT-5.4) to check for fake ground truth, score normalization fraud, phantom results, and insufficient scope. Use when user says "审计实验", "check experiment integrity", "audit results", "实验诚实度", or after experiments complete before writing claims.
Turn a refined research proposal or method idea into a detailed, claim-driven experiment roadmap. Use after `research-refine`, or when the user asks for a detailed experiment plan, ablation matrix, evaluation protocol, run order, compute budget, or paper-ready validation that supports the core problem, novelty, simplicity, and any LLM / VLM / Diffusion / RL-based contribution.