grimaldost

Craft Collection — Data skill for Claude Code

Data community

A Claude Code plugin marketplace that codifies engineering craft: disciplined Python and data-engineering practice alongside deliberate session-knowledge workflows, packaged as reusable skills.

How to install Craft Collection

This entry records only its repository, not the path inside it, so there is no exact command to give. Open grimaldost/craft-collection and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Craft Collection does

A Claude Code plugin marketplace that codifies engineering craft: disciplined Python and data-engineering practice alongside deliberate session-knowledge workflows, packaged as reusable skills.

Alternatives in Data

README

craft

[![validate](https://img.shields.io/github/actions/workflow/status/grimaldost/craft-collection/validate.yml?branch=main&style=flat-square&labelColor=2A3238&label=validate)](https://github.com/grimaldost/craft-collection/actions/workflows/validate.yml) [![license](https://img.shields.io/badge/license-MIT-39621C?style=flat-square&labelColor=2A3238)](LICENSE)

craft-collection is a Claude Code plugin marketplace with four plugins:

  • engineering-discipline — modern Python engineering standards + stack-agnostic data-engineering discipline, with mechanical enforcement (ruff/uv hooks), runnable scripts, and a self-refreshing toolchain.
  • experiment-discipline — discipline for evaluation acts: a typed record.yaml across a probe / measurement / decision tier ladder, with a validator, a derived report, small-n statistics that refuse the CLT, and a pre-registration freeze.
  • session-workflow — capture session knowledge and distill it into durable guidance, author paste-ready hand-off briefs, convene fresh-eyes review panels, behaviorally evaluate skills, keep a live inventory of the installed toolkit, sign agent-assisted work with machine-generated provenance trailers, and run a tool-dogfooding feedback loop (capture + triage).
  • humblepowers — superpowers-derived process disciplines in a calibrated register (fit-ranked dispatch, calibration-first skill authoring, TDD, root-cause debugging, brainstorming, verification, review reception, midweight planned execution, model/effort dispatch) — eval-gated, with a register linter. Replaces superpowers; never install both.

Install

**Prerequisites:** [