diff --git a/CHANGELOG.md b/CHANGELOG.md
index 4effe7c..2244620 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -28,6 +28,7 @@
 
 ### Fixed
 
+- README / `@cedarjs/pg/vite-plus` Vite+ example: `dev` is now `cedarpg run --attach --mode=dev -- vp dev` (with `cache: false`). Vite+ `dependsOn` does not forward env and task `env` only passes through the `vp` process env, so the old `command: "vp dev"` + `env: ["DATABASE_URL"]` never saw the leased URL.
 - `cedarpg run` (plain and `--attach`) and `cedarpg studio` forward `SIGINT` / `SIGTERM` / `SIGHUP` to the child. A supervisor that signalled only the `cedarpg` process (Nx, pm2, `kill <pid>`) used to leave the child running. A child killed by a signal now exits with `128 + n` (130 for `SIGINT`, 143 for `SIGTERM`) instead of 1.
 - Ephemeral host cold start no longer deletes `/dev/shm/PostgreSQL.*` and `pgserve-*`. Those are live shared-memory segments of every Postgres on the machine. Running with `CI=true` or `CEDAR_PG_EPHEMERAL_HOST=1` next to a registered local host deleted that host's segment, and every new connection to it then failed with `58P01 could not open shared memory segment` until it was restarted. Cold start now removes only cedar-pg's own leftover data dir (`/dev/shm/cedar-pg-<uid>`). The `53100` / quota hint no longer recommends `rm /dev/shm/PostgreSQL.*`.
 - TEMPLATE `cloneWorkerDatabase` reuses the worker's `<template>_c_<workerId>` in Postgres across Jest test files. The `0.2.0-beta.0` fix cached the clone on `globalThis`, but Jest resets `globalThis` and the module registry for each test file. So the second file in a worker still ran `CREATE DATABASE` and failed with `database already exists`. The in-memory memo is removed; the database is the source of truth. As a result, calling it again with a different `root` / `name` in one process no longer throws: each call clones or reuses its own name. `smoke:pg` now runs two Jest files in one worker, after a seeded crashed run.
diff --git a/README.md b/README.md
index f711f57..b2d52ea 100644
--- a/README.md
+++ b/README.md
@@ -106,15 +106,17 @@ export default defineConfig({
         env: ["DATABASE_URL", "TEST_DATABASE_URL"],
       },
       dev: {
-        command: "vp dev",
+        command: "cedarpg run --attach --mode=dev -- vp dev",
         dependsOn: ["db:acquire"],
-        env: ["DATABASE_URL"],
+        cache: false,
       },
     },
   },
 });
 ```
 
+Like Nx `dependsOn`, Vite+ `dependsOn` does not forward env, and task `env` only fingerprints and passes through variables already set in the `vp` process. So `dev` acquires once via `db:acquire`, then `cedarpg run --attach` hands the lease `DATABASE_URL` to `vp dev`. `test` needs no wrapper: the stock Vitest setup acquires and sets the URL itself.
+
 `cedarPgDev()` does not acquire. Keep `dependsOn: ["db:acquire"]`. On listen it prints a status panel (TTY, non-CI). Vite CLI shortcuts (`key` then Enter; also listed under `h`):
 
 | Key | Action                                                                 |
diff --git a/src/adapters/vite-plus.ts b/src/adapters/vite-plus.ts
index 4d01b2b..82f8344 100644
--- a/src/adapters/vite-plus.ts
+++ b/src/adapters/vite-plus.ts
@@ -20,10 +20,11 @@
  *         dependsOn: ['db:acquire-test'],
  *         env: ['DATABASE_URL', 'TEST_DATABASE_URL'],
  *       },
+ *       // dependsOn does not forward env: attach hands vp dev the lease URL.
  *       dev: {
- *         command: 'vp dev',
+ *         command: 'cedarpg run --attach --mode=dev -- vp dev',
  *         dependsOn: ['db:acquire'],
- *         env: ['DATABASE_URL'],
+ *         cache: false,
  *       },
  *     },
  *   },