Stats
Actions
Tags
From east
Create, initialise, and manage East / e3 projects end-to-end. Use when: (1) Scaffolding a new project ('create an e3 project called X', 'set up an East project', 'new East project'), (2) Choosing between an e3 project (BSL-1.1, Node + Python, durable execution) and an East project (AGPL-3.0, Node-only), (3) Driving the project lifecycle — install, build, deploy, run, watch, test — with the e3 CLI, (4) Going from an empty directory to a running dataflow.
How this skill is triggered — by the user, by Claude, or both
Slash command
/east:east-projectThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Scaffold and run East projects. This skill creates the skeleton, then you implement the logic using the **e3**, **east**, **east-ui**, and **east-py-datascience** skills.
Scaffold and run East projects. This skill creates the skeleton, then you implement the logic using the e3, east, east-ui, and east-py-datascience skills.
| Want | Use | License | Stack |
|---|---|---|---|
| Durable dataflow, workspaces, multi-runtime | e3 project | BSL-1.1 | Node + Python |
| Plain East programs, Node platform fns only | East project | AGPL-3.0 | Node-only |
The plugin ships the scaffolders on PATH as east-scaffold (Claude Code adds the plugin's bin/ to PATH on install):
east-scaffold e3 my-project # or '.' for the current directory
east-scaffold east my-project
If east-scaffold isn't found (plugin not installed in this environment), fall back to the curl bootstrap:
curl -fsSL https://raw.githubusercontent.com/elaraai/east-workspace/main/libs/east-claude-plugin/scripts/scaffold/e3.sh | bash -s -- my-project
Either way you get: package.json, tsconfig.json, pyproject.toml (e3 only), src/index.ts (a sample e3.input + e3.task), src/main.ts (package export), tests, and a Makefile.
After scaffolding, edit src/index.ts to build the user's actual logic:
e3.input, e3.task, e3.package, e3.export).ui(), Data.bind).make install # npm install + uv sync
make build # tsc
make test # export IR from TS, run Python compliance tests
make start # build → e3 repo create → workspace deploy --from-zip → dataflow run
make watch # e3 watch ./src/index.ts .repos <ws> --start (live reload)
Under the hood make start uses the current e3 CLI:
e3 repo create .repos
e3 workspace deploy .repos <ws> --from-zip /tmp/pkg.zip # imports + creates ws + deploys
e3 dataflow run .repos <ws>
e3 dataset get .repos <ws>.<name> # read a result (flat path)
east-scaffold e3 Xcd Xsrc/index.ts: define inputs + tasks implementing Y (use the e3/east skills).make install && make start (or make watch for iteration).e3 dataset get .repos <ws>.<task> to read outputs.npx claudepluginhub elaraai/east-workspace --plugin eastProvides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.