全部能力

Generate a structured paper outline from review conclusions and experiment results. Use when user says \"写大纲\", \"paper outline\", \"plan the paper\",…

Writes rigorous mathematical proofs for ML/AI theory. Use when asked to prove a theorem, lemma, proposition, or corollary, fill in missing proof steps,…

Responds to unanswered GitHub discussions and issues with codebase-informed replies. Use when clearing community question backlog.

Use when validating optimization plans through the evaluation platform with mandatory research, parallel agent evidence, sequential merge, and bounded…

Create and edit Obsidian Bases (.base files): Obsidian's native database layer for dynamic tables, card views, list views, filters, formulas, and summaries…

Draft a structured grant proposal from research ideas and literature. Supports KAKENHI (Japan), NSF (US), NSFC (China, including 面上/青年/优青/杰青/海外优青/重点), ERC…

Update user-facing documentation when code changes. Use when asked to update docs, review docs, handle documentation changes, run scheduled documentation…

Audits test maintainability and structure: layout drift, orphan tests, fragmented duplicates, test-to-source mapping, and domain grouping. Use when auditing…

Generate publication-quality AI illustrations for academic papers using Gemini image generation. Creates architecture diagrams, method illustrations with…

Researches standards and patterns via MCP Ref for Story Technical Notes. Use when Story needs standards research before task generation.

Compile LaTeX paper to PDF, fix errors, and verify output. Use when user says \"编译论文\", \"compile paper\", \"build PDF\", \"生成PDF\", or wants to compile LaTeX…

Autonomous research review loop using any OpenAI-compatible LLM API. Configure via llm-chat MCP server or environment variables. Trigger with "auto review loop…

Blog strategy development including topic cluster architecture with hub-and-spoke design, audience mapping, competitive landscape analysis, AI citation surface strategy across ChatGPT/Perplexity/AI Overviews, distribution channel planning (YouTube, Reddit, review platforms for GEO), content scoring targets, measurement framework, and content differentiation through original research and first-hand experience. Use when user says "blog strategy", "content strategy", "blog positioning", "what should I blog about", "blog topics", "content pillars", "blog ideation".

Researches competitive benchmarks and generates optimization hypotheses for identified bottlenecks. Use after profiling.

Profiles runtime performance with CPU, memory, and I/O metrics. Use when measuring bottlenecks before optimization.

You must use this when producing any research prose — literature reviews, syntheses, analyses, methodology descriptions, discussion sections, abstracts, or any…

Workflow 1 adaptation for robotics and embodied AI. Orchestrates robotics-aware literature survey, idea generation, novelty check, and critical review to go…

Generate a conference poster (article + tcbposter LaTeX → A0/A1 PDF + editable PPTX + SVG) from a compiled paper. Use when user says \"做海报\", \"制作海报\",…

Run Karpathy-style autonomous ML search on a real training repository. Use when the user needs to set up or operate `karpathy/autoresearch`, choose the right run mode (setup, `program.md`, bounded loop, result interpretation, or constrained-hardware adaptation), and preserve the immutable `prepare.py` / 300-second / `val_bpb` contract. Not for prompt evaluation, LLM app observability, or repo-local `SKILL.md` optimization — route those to LangSmith, Promptfoo, Braintrust, or `skill-autoresearch`. Triggers on: autoresearch, autonomous ML experiments, `program.md`, `train.py`, `val_bpb`, overnight GPU loop, fixed eval harness.

Researches best practices and creates documentation (guide, manual, ADR, or research). Use when project needs standards-based documentation for a topic.