noWorkflow skill
A skill that teaches an agent to run Python scripts under noWorkflow, compare trials and trace where a result value came from.
- 130 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- `now run` executes your script and stores file copies and environment variables, secrets included, in `.noworkflow/`.
- `now push` sends the store to a server together with the trials.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/noworkflowDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- `now run` executes your script and stores file copies and environment variables, secrets included, in `.noworkflow/`.
- `now push` sends the store to a server together with the trials.
This is third-party code. Review the repository files before installing.
What it does
noWorkflow records the provenance of a Python script without any code changes. `now run script.py` stores a trial: the script and imported modules, arguments, environment, function calls with arguments and return values, variable evaluations and every file the script opened, before and after. Data live in a `.noworkflow/` directory next to the script. The `skills/noworkflow/SKILL.md` skill tells an agent the commands: listing and showing trials, comparing two runs, finding input files and functions a result depends on, exporting to Prolog, PROV-N, Graphviz and notebooks, restoring an earlier version of the script and files, and opening the visualization. Known bugs are cited with repository issue numbers.
Who it is for. Researchers and data engineers reproducing computations who want to ask an agent what changed between yesterday's and today's run.
Good fit when
- After editing a script you need to see how the new run differs from the last one.
- An output file looks doubtful and the agent traces the inputs and functions behind it.
- The script and data of the third run must be restored from `.noworkflow/`.
- Result provenance is exported to PROV-N or Graphviz for a paper.
Not a fit when
- The script is not Python, and noWorkflow records external programs only as launches.
- The store is about to go to colleagues unchecked, though it holds the user name, hostname and environment variables.
- You need production service monitoring, while the tool is aimed at scientific experiments.
Example request
Run `simulation.py` under noWorkflow and tell me what changed since the last run.Limitations
The script runs under the interpreter where noWorkflow is installed, so the package goes into the same environment as the script's dependencies. The `.noworkflow/` directory keeps copies of every file the script read or wrote and every environment variable, tokens and passwords included if they sit there, plus the user name and hostname. Check the store before sharing it and keep `.noworkflow` out of public repositories. Rendering graphs needs the system Graphviz `dot` program. `now import` is missing from the help list (issue #253); other limits are listed in the skill.
How to disable. Delete `~/.claude/skills/noworkflow` and, if you no longer need the package, remove `noworkflow` from the Python environment. `.noworkflow/` directories next to scripts are deleted separately.
FAQ
How do I record a run?
`now run script.py` executes the script and stores a trial. `now list` shows them and `now <command> -h` gives help for any command.
Where is the data stored?
In `.noworkflow/` next to the script: `db.sqlite` and a `content/` directory with file version copies. Other commands look for the store in the current directory or in `--dir`.
What does the `[all]` install add?
IPython, Jupyter, pyswip-alt, matplotlib and numpy.
Is Graphviz required?
Only to render graphs from `.dot` files. It is a system package, not a Python dependency.
How do I open the visualization?
With `now vis`: it starts an interactive Flask-based panel that `pip install noworkflow` already installs.
Discussion
No questions or reviews yet.
Related
SkillsExcel analysis skill that profiles a sheet, cleans it, reconciles totals against the sheet's own sums and then releases numbers.
Probabl skills for machine learning experiments: a scikit-learn and skrub pipeline, evaluation, audit and a link to Skore Hub
A skill and Python package for hyperparameter tuning: the agent reads the code, proposes the next set of values, and your own run measures the result
An agent skill that turns an arxiv paper into a minimal Python implementation, with code anchored to sections and equations and gaps flagged honestly
Around 200 skills for Claude Code on bioinformatics, chemistry and biostatistics, from RNA-seq analysis to molecular docking
17 skills for computational biology: RNA-seq, single-cell, protein structure, docking, literature and figures
README in short
Universities UFF and NYU have developed the project since 2013. Releases appear on PyPI, and the README covers commands, visualization and use through an AI agent.