Agent Skills by kittycafe
Showing 41 vetted skills indexed across 1 GitHub repositories.
code-writing
Enforce consistent coding standards for generation, refactoring, and modification tasks.
code-review
Diagnose architectural and performance issues in user-submitted code.
paper-write
Draft a complete LaTeX paper from an outline and narrative.
paper-poster
Extract research content from a compiled paper into a conference poster.
proof-writer
Draft rigorous mathematical proofs for ML/AI theory claims.
formula-derivation
Constructs structured derivation outlines from research problem statements and invariant objects.
training-check
Monitor WandB metrics periodically to detect training issues like NaN/Inf or stagnation.
result-to-claim
Judge experimental results against intended claims and route pivot, supplement, or confirm actions.
mermaid-diagram
Generate Mermaid diagrams from user requirements and verify syntax.
research-refine-pipeline
Orchestrate an end-to-end research-to-proposal workflow with standardized outputs.
auto-review-loop-llm
Automate multi-round research review with OpenAI-compatible LLMs and persistent logs.
experiment-plan
Convert research proposals into structured experiment plans with run orders and budgets.
idea-creator
Generate and rank research ideas with hypotheses and pilot plans.
grant-proposal
Draft structured grant proposals with agency-specific sections and timelines.
monitor-experiment
Monitors remote experiment sessions and aggregates their latest results via SSH.
research-pipeline
Automate end-to-end research workflows from idea discovery to submission-ready manuscripts.
research-refine
Refine vague research directions into concrete, phase-driven research plans.
ablation-planner
Generate an Ablation Plan with component changes and hyperparameter sweeps.
idea-discovery
Chain literature survey, idea generation, novelty checks, and critique into an automated workflow.
paper-compile
Compile LaTeX papers to PDF and fix common compilation errors.
paper-slides
Generate Beamer PDF slides and editable PPTX from a compiled paper.
analyze-results
Parse JSON/CSV ML results to compute statistics and generate comparison tables.
paper-figure
Generate publication-ready plots, multi-panel figures, and LaTeX tables from experimental data.
run-experiment
Deploy and run ML experiments on local or remote GPU servers.