More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Citable Python Packages from Research Code

    1

    Turn research scripts into a pip-installable, citable Python package: src/ layout with uv, ruff and pytest, CITATION.cff and a How to cite section, a lock file for reproducibility, and your code moved without rewrites.

    Free

    2 usesSecurity scanned
    Citable Python Packages from Research Code

    Citable Python Packages from Research Code

    Example session with this skill installed

    Turn my two analysis scripts, fit_arrhenius.py and read_runs.py, into a pip-installable package I can cite in my thesis. There's a small example dataset in data/rates.csv. Author: Alex Rivera (ORCID 0000-0002-1825-0097), MIT licence, no paper yet. I'd like an examples folder. Go ahead without waiting for my confirmation, run the tests at the end, and flag anything in my code you'd change instead of changing it.

    • Read your context and instructions
    • Compiled the citable python packages

    Package arrhenius-fit (import arrhenius_fit), built from the two scripts:

    • src/arrhenius_fit/: both scripts, unchanged except one import made relative; functions re-exported from __init__.py
    • pyproject.toml (hatchling, numpy>=1.24, matplotlib>=3.7, dev group with pytest, ruff, pyyaml), uv.lock, .python-version
    • CITATION.cff with the ORCID, LICENSE (MIT), README.md with quick start and "How to cite" (BibTeX)
    • examples/01_fit_example_data.py (prints Ea = 52.7 ± 0.5 kJ/mol, saves a plot) and a test that runs it

    Checks: uv run pytest: 10 passed. uv build: wheel and sdist built. Ruff findings only in the two moved files, left for you.

    Flagged, not changed

    • min_replicates counts usable rows, not replicates per temperature
    • fit_arrhenius fails on plain lists (1.0 / T)
    • hard-coded cluster path in the __main__ block
    • matplotlib is a required dependency only because of the plot in __main__

    Still to do: repository URL placeholder, no DOI yet, git init, PyPI not configured.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    What it does

    Your agent turns loose research code, or just an idea, into a Python package other people can install, run and cite. It:

    • Builds a clean package with a fixed, modern toolchain: uv for the environment and lock file, hatchling to build, ruff to lint, pytest to test, and a src/ layout so tests run against the installed package.
    • Makes it citable: a CITATION.cff (authors, ORCID iDs, licence, version), a README with a "How to cite" section and BibTeX, and a test that fails if the citation is invalid, still has placeholders, or drifts from the package version.
    • Records what your results ran on: lower bounds in pyproject.toml, exact versions in a committed uv.lock, the Python version in .python-version.
    • Keeps your code yours: existing scripts are copied into the package verbatim and the originals are left in place. The only edit is turning imports between your own scripts into package imports. Code that runs on import, hard-coded paths, undeclared imports and lint findings are flagged for you to decide, never silently rewritten.
    • Runs the checks (uv sync, ruff, pytest) and hands over the exact commands plus a list of what is still a placeholder.

    Optional extras

    On request: an examples/ folder with a test that runs every example, a notebooks/ folder, a GitHub Actions CI workflow testing the oldest and newest Python, and a Dockerfile that installs the locked versions. When you want a DOI, it prepares the release and gives you a checklist for Zenodo's GitHub integration, with a Code availability statement for your paper.

    What's included

    • SKILL.md: the workflow for both starting points (an idea, or existing scripts)
    • references/: annotated templates for pyproject.toml, CITATION.cff and its test, the README, CI, Docker and examples, plus guides to reproducibility and releases with a DOI

    Part of the Aurora Lab research pipeline

    Works entirely on its own. If you keep a research/ folder, it can read research/refs.bib and research/paper/ to fill in the paper citation. No skill calls another.

    Python only, with a fixed toolchain (uv, hatchling, ruff, pytest); it does not set up Poetry, conda or setuptools projects. It packages your code but does not judge whether the science is right: the tests it writes are smoke tests on small known inputs, not a validation of your method. It does not create the GitHub repository, publish releases, mint a DOI or upload to PyPI; it prepares the files and gives you the steps. Zenodo and GitHub Actions details reflect their documentation as of October 2026; check the current guides when you release.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    2 uses

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 29 days ago

    • Free with an Agensi account

    Requires

    Python 3.10+
    Internet access
    Python
    uv (Python tool runner)
    cffconvert
    Git

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access

    File Scopes

    src/**
    tests/**
    pyproject.toml
    uv.lock
    CITATION.cff
    examples/**
    research/refs.bib

    Reads the scripts you point it to. Writes the package files in your project: pyproject.toml, uv.lock, .python-version, .gitignore, src/, tests/, README.md, CITATION.cff, LICENSE, and on request examples/, notebooks/, a CI workflow and a Dockerfile. It copies your code into the package without rewriting it and deletes nothing: your original files stay where they are. It runs uv, ruff and pytest locally. Network access is only for uv sync, which downloads the package's dependencies from PyPI (and, if needed, a Python interpreter) the first time; the agent asks before the first sync. It never contacts GitHub, Zenodo or PyPI to publish anything: those steps are a checklist for you.

    Listed29 days ago

    What's inside

    Frequently Asked Questions

    No. Existing scripts are copied into the package verbatim, and the originals stay where they are. The one edit it makes is turning imports between your own scripts into package imports (from load import x becomes from .load import x), and it tells you about it. Everything else it would change (code that runs on import, hard-coded paths, undeclared imports, lint findings) is listed for you to decide. It runs ruff --fix or ruff format on your code only if you agree.

    Python 3.10+ and uv. The package's own dev tools (pytest, pytest-cov, ruff, pyyaml) are installed by uv into the project's .venv. Without uv, the agent still writes every file and gives you the commands to run later.

    Only for uv. The first uv sync downloads your package's dependencies from PyPI (and a Python interpreter if you don't have the one requested). The agent asks before that first sync. The skill itself contacts nothing.

    It prepares everything for one: version bump, CITATION.cff, release notes, README badge placeholder, and a step-by-step checklist for Zenodo's GitHub integration, plus a Code availability statement for your paper. You publish the GitHub release and switch on the repository in Zenodo yourself; the repository must be public.

    No. Once the repository is on GitHub, people can install a tagged release straight from git (pip install "name @ git+<repository URL>@v0.1.0"). PyPI publishing is an optional later step, which the handoff points out.

    No. It writes smoke tests that call each function on a small input with a known answer, to show the package works after the move, plus the citation test. Whether your method is correct is outside its scope.

    Yes. Describe the package in a sentence and its first call; it scaffolds the same structure with stub functions and tests, so the project installs and its tests run before any real logic exists.

    Any agent that loads SKILL.md skills, including Claude Code, Codex CLI, Cursor and Gemini CLI. It needs a terminal to run uv, ruff and pytest.

    New in Research & Analysis