marpogaus-skills

Plugin marpogaus-skills from marketplace marpogaus-skills in MArpogaus/agent_skills. Commands read from the repo's manifest — 1 month ago.

Skills (12)

• auto-hpo-studies

  Autonomous hyperparameter optimization loop: define target and search space, then iterate update-config run monitor log commit until convergence. Uses MLflow for tracking, sleep-based polling for long-running tasks. Use when the user asks to tune hyperparameters autonomously. requires: - ml-project

• create-skill

  Create a new agent skill (skills/<name>/SKILL.md) in this collection: determine requirements, domain, workflows and conventions, write the SKILL.md, and register it. Use when the user asks to add, write or scaffold a skill. stack: meta

• dev-conventions

  How the user and the agent work together on software projects, in any language: design principles (simplest thing that works, no reinvention, no backwards compatibility before 1.0), when to ask and when to act, verification by running and by looking, the continuous review loop (a commit monitor star

• emacs-config

  Conventions and architecture knowledge for Marcel's literate Emacs configuration (MArpogaus/emacs.d): org-only editing workflow, naming schemes, package idioms, the my/pycell notebook layer, header-line buttons, display-property rendering lore, and the repo's git flow. Use when working in that confi

• emacs-package

  Conventions for building, checking and publishing Emacs Lisp packages the way MArpogaus does: repository layout, file headers, the make based check pipeline, GitHub Actions matrix, ERT tests, and the MELPA submission steps. Use when creating a new package, extracting one from a configuration, revie

• git-conventions

  Git workflow conventions: two-branch flow, conventional commits, pre-commit hooks, dependabot, tagging, releases, and the push and attribution rules for agents. Load before any commit, branch, push, pull request, tag or release, and when setting up hooks or CI. requires: [] stack: git

• latex-paper-revision

  LaTeX paper revision workflow: structural analysis, content generation, language polishing, figure/table integration, and final cleanup. Use when revising an existing draft; a new paper follows paper-writing. stack: latex

• ml-project

  Research ML project conventions: DVC pipeline orchestration, MLflow experiment tracking, HPO study logs, paper-oriented code structure, per-target model configs, and reproducible figure generation. Use when setting up or changing the experiment code of a research project. requires: - git-convention

• org-beamer-presentation

  Author Beamer talks as Org mode files exported to LaTeX and PDF, in the denote-named presentation style: three interchangeable designs (metropolis-dark, Szeged/dolphin, SimplePlus), a bare shared header with opt-in extras, H:2 section/frame structure, column and overlay layouts, biblatex citations,

• paper-writing

  End-to-end academic paper writing workflow for probabilistic ML research: LaTeX project scaffolding, figure generation, writing with IEEE/Elsevier/UAI templates, revision cycles, and venue submission. Use when writing or submitting a paper; the experiment code follows ml-project, and a revision of a

• python-dev

  Python project conventions: setuptools, ruff I/E/F/D/UP, NumPy docstrings, MkDocs + mkdocstrings + mike docs, pytest xdist, uv, commitizen, matplotlib figure patterns, and skeleton-based project setup. Use when creating, changing or reviewing a Python project. requires: - git-conventions stack: pyt

• repo-tour

  Guided step-by-step code tour of an unfamiliar repository: orient cheaply, link one entry point, stop, then answer questions and offer the next stop. One stop at a time, markdown file links with line anchors, claims verified by running the code. Use when the user asks for a tour or walkthrough of a

Similar plugins