lopebase 0.0.0-stage → 0.1.0

package/README.md

@@ -1,3 +1,28 @@
-# Temporary Holding Version
+# lopebase
 
-This version is a temporary placeholder for this package. An operational version to replace this has been submitted for review and is awaiting a staged release.
+Set up a project in LopeBase from its own repo. Your coding agent does the steps
+and stops for you only when it has to. See ADR 0020.
+
+```sh
+npx lopebase          # scan, pick what to connect, hand off to your agent
+npx lopebase login    # connect this computer: approve a code in the browser
+npx lopebase scan     # what the scan found (--json for machines)
+npx lopebase prompt   # print the agent prompt instead of launching an agent
+npx lopebase mcp      # stdio MCP server that forwards to your workspace
+```
+
+`login` uses the device flow: it shows a short code, opens your workspace, and
+waits while you press Connect. `login --paste` takes a token from Settings,
+Agent access instead, for workspaces that do not offer device sign-in.
+
+From this monorepo: `bun run lopebase` (from the repo root). To publish:
+`bun run build` makes `dist/lopebase.js` for Node, and `npm publish` runs
+verify and build first.
+
+The scan reads package.json files, config file names, and env var names from
+example env files. It never reads a real `.env` file and never sends a value.
+The token is saved to `~/.config/lopebase/auth.json` (mode 600), or read from
+`LOPEBASE_TOKEN` and `LOPEBASE_URL`.
+
+`src/detect.ts` is also imported by the backoffice's `detect_stack` tool, so the
+terminal and the server give the same answer.