How can I contribute?
Templates, specialists, connectors and skills are written by researchers. This page explains why that matters, how each kind is added, and how you are credited.
Why contribute
Why contribute to E2ER?
A template, specialist, connector or skill written by someone who knows a method well makes every study that uses it better. For instance, a skill on the Gioia method written by a qualitative researcher is what the specialists read when they code interviews. The part keeps your name, and every study that uses it credits you.
How am I credited?
Each part has its own page with its authors, their ORCID iDs and a version. Every study built with it lists it in its dossier, the page the paper links to in its first-page footnote, pinned to the exact version used. The Insights tab of the part and of your profile shows which studies use it. The site shows no scores or rankings of people.
Can others cite what I contribute?
Yes. Each listed version has a stable address, and the dossier of a study records which version it used. A DOI for each version, deposited on Zenodo, is planned.
What happens when my part changes later?
A new version is added and the old one stays readable. Studies keep the version they were produced with, because their dossiers pin it by content hash.
How, for each kind
Templates see all
What is a template?
A research process written as a list of steps in a text file: which specialists run, in which order, and which checks must pass before the run continues. The empirical template is the one E2ER runs by default.
How do I write one?
Copy pipelines/empirical.toml, change the steps and give the file a new name. A step either hands a phase to the strategist, runs a check, runs a fixed set of specialists, or collects reviewers' verdicts. The file is validated against the template schema, and a run can be started with it by name.
Do I need to program?
No. A template is a text file. It can only use specialists and checks that exist; a new specialist is a separate contribution.
Specialists see all
What is a specialist?
An AI role with one task and one defined result, for instance the identification strategist, who writes identification_strategy.md. A run accepts the result only when it meets the specialist's contract.
How do I add one?
Add its name and the file it produces to src/core/specialists/registry.py, list the skills it reads in src/skills/loader.py, and add a test entry in tests/conftest.py. CONTRIBUTING.md in the repository describes each step.
I work in a field the theory lab covers. What can I do?
Edit your field's specialist. Its diagnostic question, canonical models and known biases should come from people who work in the field. Open the specialist and choose "Edit this specialist"; you are credited on it next to the original authors.
Connectors see all
What is a connector?
Anything a run reaches outside your machine: an AI provider, a literature or data source, another pipeline, a place to publish. Every connector states what leaves your machine and what you must provide, such as an API key.
How do I add one?
A data source is a module under src/modules/data that describes itself and fetches on request; the FRED and Yahoo Finance modules are short examples to copy. An AI provider implements one method and is registered in src/modules/llm/registry.py. Say in the description exactly what the connector sends out.
Can I connect another research pipeline?
Yes. A pipeline can run as one step of a template. The coarse referee report is listed this way as a planned connector, with its author, licence and what it sends to its model.
Skills see all
What is a skill?
A written guide that specialists read before they work, for instance on difference-in-differences with staggered adoption or on writing a literature review. It is plain text; the specialists receive it as part of their instructions.
How do I add one?
Write a markdown file with a title, a one-paragraph description and sections with rules, checklists or examples, and add it under skills/files/. Then name the specialists that should read it. No code changes are needed.
Can I keep my skills in my own repository?
Yes. Publish them in your repository with a licence and list the pack in the RISE catalogue. E2ER shows them with your name and links to your repository.
Review, licence and data
Review process proposed
Who reviews a contribution?
Proposed: machine checks first (the description is complete and the part runs on sample material), then at least one reported evaluation, then a reader who knows the method and a maintainer. Each step accepts, returns with reasons, or declines with reasons. Until these rules are settled with contributors, contributions go through pull requests on GitHub.
Which licence should I choose?
MIT, Apache-2.0 or CC BY 4.0 let E2ER ship the part with the software. A part under CC BY-NC can be installed by users but not redistributed with E2ER.
Do I have to share data or a paper to contribute?
No. A template, specialist, connector or skill is published on its own. Studies are published separately, from your own machine or repository.