"Use when a coding agent edits the Dagster OSS repository itself:
Scanned 9/8/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill repo-development --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Repo Development?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-repo-development-57e7b534)More formats (shields.io, HTML) on the badges page.
---
name: repo-development
description: "Use when a coding agent edits the Dagster OSS repository itself:
package lookup, Python/UI/docs validation, coding conventions, mandatory ruff,
safe test selection, CLI entry point reinstall notes, and git or stack
constraints."
disable-model-invocation: true
metadata:
disco-role: operating
license: Apache 2.0
---
# Repo Development
Use this sub-skill when changing the Dagster OSS monorepo rather than using Dagster as an application framework. It routes future agents to the repo-specific development workflow, package map, validation commands, style rules, and safety constraints needed for edits in this checkout.
## Route First
- Use this sub-skill for locating Dagster packages, choosing validation commands after repo edits, applying Python style conventions, docs checks, UI checks, or repository contribution workflow rules.
- Route public `dagster` CLI usage and user-project local development to `../cli-local-development/SKILL.md` if that sub-skill exists.
- Route GraphQL API and webserver runtime behavior to `../graphql-and-webserver/SKILL.md` if that sub-skill exists.
- Route deployment operations, production instances, and daemon/run-launcher behavior to `../deployment-operations/SKILL.md` if that sub-skill exists.
- Do not use this sub-skill for broad integration-library internals, Dagster Cloud, release automation, Buildkite deep internals, or UI implementation details beyond validation workflow routing.
## Start Here
1. Identify changed paths and package ownership before editing; package paths and validation commands differ across core Python, libraries, webserver, GraphQL, Pipes, docs, and UI workspaces.
2. After every Python file edit, run `make ruff` from the repository root before considering the task complete.
3. Use `uv` for Python package management; if command-line entry points change in a `setup.py` or package metadata, reinstall the affected package with `uv pip install -e .` from that package root.
4. For UI or GraphQL schema changes, run UI checks from `js_modules/`; schema changes require `make generate-graphql` before `yarn tsgo` or `yarn lint`.
5. Prefer focused tests that match changed files before broad suites; avoid credentialed, networked, Docker-heavy, or expensive tests unless the user explicitly asks.
6. Never push directly with `git push`; ask first. When stack metadata matters, inspect with `gt log` before manual git inference.
## References
- [Repo development workflows](references/workflows.md) for package locations, command selection, Python/UI/docs validation, style conventions, test strategy, and git constraints.
- [Troubleshooting](references/troubleshooting.md) for install/import failures, optional dependency gaps, CLI/API misuse while developing the repo, UI/docs validation failures, and workflow-specific recovery steps.
- [Validation command selector](scripts/select_validation_commands.py) for a safe helper that suggests commands from changed paths without executing them.
## Safety Notes
- Treat `make ruff` as mandatory after Python edits even if the changed file is small or generated.
- Do not search Python code under `.tox` directories; they are temporary environments and can mislead package ownership or dependency conclusions.
- Search package dependencies in `setup.py` files as the source of truth for this repo.
- Do not run broad native examples, Docker-backed tests, or external-service integrations as routine validation without confirming scope and cost with the user.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!