ultimatile

Single source of truth for universal code-quality items. Definitions live in items/<slug>.md; audit and preflight skills reference items by slug.

Full review pipeline from local changes through PR, CodeRabbit review, postmortem elevation, and umbrella drift join. Pauses at the user-controlled merge gate between Phase 4a and 4b.

Full review pipeline from local changes through PR, Copilot review, postmortem elevation, and umbrella drift join. Pauses at the user-controlled merge gate between Phase 4a and 4b.

Preflight sweep of quality-list items before or during implementation, framed as 'what to set up so done-check passes at the end'. Dual of done-check.

Create a GitHub PR and poll until the auto-triggered CodeRabbit review arrives.

Investigate a GitHub issue or free-text task with empirical (subagent probes) and derivational (in-context deduction) hypotheses, producing a vetted implementation plan with Inconclusive / Deferred items. Accepts an issue number or task text.

Narrow Codex pass on a newly added contract test, verifying it is not tautological and would fail on the original buggy implementation. Lightweight alternative to the full codex-review / copilot-review loop.

Pre-PR code review on the current branch using the OpenAI Codex CLI, with an iterative fix loop.

Create and update paper reading notes in the user's Logseq vault. Use when summarizing a paper, recording results from a paper reading session, or appending notes to an existing paper page. The vault is shared between Logseq (daily-use viewer) and Obsidian (CLI access). Writes use Logseq syntax (key:: value, tab-indented outline), reads can use Obsidian CLI for backlinks/search or rg for raw content.

Write and review Japanese prose (articles, long-form docs) free of LLM rhetoric-excess, with consistent register and disciplined terminology. Universal layer — delegates project-specific house style to the project's CLAUDE.md / GLOSSARY.md. Use when drafting or reviewing Japanese articles, blog posts, or documentation prose.

User's opinionated Julia preferences. Four real frictions: (1) ASCII-only identifiers — never use Unicode (π, ≤, χ, …) because editor / terminal glyphs collapse confusables (e.g. x vs χ) and the resulting bugs are silent; (2) Tooling — scaffold with `jtc create MyPackage.jl`, manage deps with `mise run add`, and use `TestItemRunner.jl` with tags for test selection (mise tasks filter by tag); never `Pkg.generate` / `Pkg.add` / `Pkg.test` by hand; (3) Union{T, Nothing} for optional / fallible values rather than throwing or sentinel returns; (4) parameter-pack factory pattern — @kwdef struct with parametric types fed by a factory constructor that takes a loose Dict{String, Any} from CLI / config parsing. Plus opinions on multiple dispatch, type stability, performance idioms (@view, @inbounds, @simd, StaticArrays), and Project.toml compat. TRIGGER when writing or editing any Julia (.jl) file, Project.toml, test/, or mise.toml. SKIP for non-Julia work.

User's opinionated Python preferences. The two non-negotiables are (1) ALWAYS use uv for package management — never pip/poetry/conda/venv directly — and (2) ALWAYS add type annotations to every function signature. Plus secondary preferences (ruff, pyright strict, pydantic v2, 3.10+ syntax) and known pitfalls. TRIGGER when writing or editing any Python (.py) file, pyproject.toml, requirements.txt, or any package-management or environment setup. SKIP for non-Python work.

Convert arXiv papers to Markdown documentation. Fetches available materials from arXiv (LaTeX source when available + PDF), converts LaTeX to Markdown via pandoc (happy path). PDF-only papers get a naive single-column fallback — use the specialized PDF scripts for better results.

Look up arXiv paper metadata via the arXiv API. Use when you need to get a journal DOI from an arXiv ID (for OpenAlex integration), or find an arXiv ID from a title/keyword search (for arxiv-doc-builder). Requires the `arxiv` Python package.

HPC cluster workflow automation - sync files to remote cluster, submit scheduler jobs (Slurm/PJM), and monitor job status. Use when working with crewster.toml, submitting HPC jobs, syncing to clusters, or checking job status.

Search CUDA-X library documentation (cuBLAS, cuTENSOR, cuTensorNet, cuSOLVER, etc.) to find API symbols, functions, and types. Use when you need to look up CUDA library APIs, discover available functions, or find documentation URLs for specific operations.

Generates commit messages in Conventional Commits format based on staged changes (git diff --staged). Use this skill whenever the user asks to generate a commit message, create a commit, write a git commit, or format a commit. Also trigger when the user says things like "create a commit message", "what should my commit message be?", "help me commit this", or mentions "conventional commits". Even if the user just says "commit" without additional context, consider using this skill to produce a well-formed message.

Survey citing or referenced papers for a given seed paper using OpenAlex API. Use when the user asks to find follow-up papers, citing papers, references, or related literature for a specific paper identified by DOI or arXiv ID. Also use when asked to check prior work, search for related publications, or build a literature survey around a specific paper.

Create lazygit custom commands and keybindings. Use when the user wants to add, modify, or debug lazygit custom commands, including YAML config entries and associated shell scripts. Triggers on mentions of "lazygit custom command", "lazygit keybinding", or lazygit config.yml custom command editing.