what-i-do
Skill1 starsChanged 14 days agoMITView on GitHub
Orientation map and canonical catalogue of the mad-science data science skills. Use when the user asks how the skills fit together, which command to run next, where to start an experiment, or needs a quick overview of the hypothesis-to-notebook flow.
In this repository15 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- llms.txtllms.txt
- critic-reviewSkill
- developing-data-science-notebooksSkill
- expectation-driven-scienceSkill
- experiment-graphSkill
- experiment-pipelineSkill
- exp-reviewSkill
- hypothesysSkill
- let-it-ripSkill
- literature-researchSkill
- mad-science-setupSkill
- notify-gchat-human-waitSkill
- to-articleSkill
- to-docSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 27
Every implementation is reviewed by /exp-review to ensure the code state is correct before moving on.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
What I Do
This skill is the orientation map for the experiment-driven workflow — and the canonical catalog of every skill in this repo. It explains how the data-science commands fit together, tells you which one to invoke next, and lists every skill with its role. It does not implement anything itself — it routes.
Catalog governance: every skill in this repo must be registered here, below, before it is considered final. If a skill changes the experiment flow, the Main Flow and/or Flow Checks sections must be updated to reflect it. The
README.mdof the repo mirrors this file; this file wins on divergence.
Main Flow: Hypothesis → Notebook Experiment
You have a hypothesis and want to test it. The flow has four steps, in order:
/hypothesys— Understand and sharpen the user's hypothesis through a short interview. This is the first step: it clarifies the hypothesis, the premises, and the goal metrics of the experiment before anything is built./to-doc— Write the experiment doc for the idea to be developed. This fixes and consolidates the ideas, giving the next step a stable contract to implement against./experiment-pipeline— Transform the doc into a pipeline workflow, breaking it into small slices to run and tracking the experiment rounds./let-it-rip— Implement the hypothesis safely in a notebook.- Follows the rules of
/expectation-driven-scienceso the science code stays rigorous (expectation before trust, vertical slices, deviation-as-information). - Every implementation is reviewed by
/exp-reviewto ensure the code state is correct before moving on.
- Follows the rules of
Standalone
Skills used on their own — independent from the main experiment flow, not a step in the hypothesis → notebook pipeline.
| Skill | Role | When to use |
/literature-research | Deep research on high-trust primary sources; saves findings as Markdown | State of the art check, next to/after step 1 |
/to-article | Turns experiment results/docs into a scientific-technical article | When results are ready to publish |
/mad-science-setup | Configures repo layout and documents | Once, before first use of other skills |
/notify-gchat-human-wait | Sends Google Chat card when agent waits for a human | Automation/operations (outside science flow) |
/critic-review | Reviews the whole project — docs, experiments, results — against the state of the art and flags gaps, weak evidence, wrong assumptions | At the end of the project, or when you want a critical second look |
/experiment-graph | Renders a read-only interactive graph of the pipeline doc — status, dependencies, critical path | Any time you need a visual overview of docs/experiments-pipeline/*.md |
/what-i-do | This orientation map and canonical skill catalog | Whenever you don't know which skill to invoke next |
Principles
Development principles — reference material that guides how to develop, not steps of the flow.
| Skill | Role | When to use |
/expectation-driven-science | Hypothesis → prediction → verification discipline | Any analysis/model driven by explicit expectations |
/developing-data-science-notebooks | Production-grade notebook practices (structure, modularization, reproducibility) | Creating/refactoring/reviewing notebooks |
Flow Checks
| Phase | Goal | Output | Prevents |
/hypothesys | Clarity of the question | Hypothesis + premises + goal metrics | Undefined falsifiability (hindsight) |
/to-doc | Experiment contract | Doc with baseline, metrics, validation | Trust before expectation |
/experiment-pipeline | Slice-based execution | Sliced pipeline + tracked rounds | Untracked experiment rounds |
/let-it-rip | Rigorous implementation | Notebook + expectation-driven rules | Ad-hoc science code |
/exp-review | Final verification | Approved code state | Unreviewed error / bias |
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.