Agent Skills by Li Zixuan
Showing 39 vetted skills indexed across 1 GitHub repositories.
THUFanZdpaper-write
Generate LaTeX paper sections and BibTeX bibliography from an outline and narrative plan.
THUFanZdpaper-poster
Generate a 4-column IMRAD conference poster with LaTeX source from a compiled paper.
THUFanZdproof-writer
Construct and verify mathematical proofs for ML/AI theory.
THUFanZdformula-derivation
Derive structured research formulas from problem statements with classified steps.
THUFanZdtraining-check
Monitor WandB metrics to detect NaN, divergence, or idle GPUs.
THUFanZdresult-to-claim
Judges whether experimental results support claims and routes to next steps.
THUFanZdmermaid-diagram
Generate Mermaid diagrams from requirements and verify syntax with Mermaid CLI.
THUFanZdresearch-refine-pipeline
Transforms vague research directions into final proposals and experiment roadmaps.
THUFanZdauto-review-loop-llm
Automate multi-round research review and fix iterations via OpenAI-compatible LLM APIs.
THUFanZdexperiment-plan
Convert research proposals into claim-driven experiment plans with run order and budgets.
THUFanZdidea-creator
Rank actionable research ideas from a broad direction into top candidates with validation plans.
THUFanZdgrant-proposal
Convert research ideas into structured grant proposal drafts for multiple agencies.
THUFanZdmonitor-experiment
Monitor running experiments and aggregate logs, JSON outputs, and summaries.
THUFanZdresearch-pipeline
Automate the research lifecycle from idea discovery to submission-ready paper.
THUFanZdresearch-refine
Codify a strategic problem and run staged reviewer rounds with drift checks.
THUFanZdablation-planner
Design ablation studies with component removals, hyperparameter sweeps, and execution order.
THUFanZdidea-discovery
Orchestrate a structured workflow from literature review to refined proposal with experiment plan.
THUFanZdpaper-compile
Compile LaTeX documents into submission-ready PDFs with latexmk and automatic error fixes.
THUFanZdpaper-slides
Convert compiled research papers into Beamer slide decks with speaker notes.
THUFanZdanalyze-results
Parse JSON/CSV ML experiment outputs into structured comparison summaries with statistics.
THUFanZdpaper-figure
Generate publication-quality figures and LaTeX-ready tables from experimental data.
THUFanZdrun-experiment
Deploy and run ML experiments on local or remote GPU servers.
THUFanZdauto-review-loop
Automate multi-round research reviews with a secondary Codex agent.
THUFanZdpaper-plan
Generate a paper outline from review conclusions and experiment results.