fix(ci): sequential steps + single install to prevent OOM on runner

Each step was re-running pnpm install independently, and all quality
steps (typecheck, lint, format, test) ran in parallel. On merge commits
with more accumulated code this pushed the CI runner over its memory
limit (exit code 254 = OOM kill).

Fix:
- install once, share node_modules via Woodpecker workspace volume
- sequential execution: install → typecheck → lint → format → test → build
- corepack enable in each step (fresh container) but no redundant install

.woodpecker/ci.yml

@@ -1,22 +1,25 @@
 variables:
   - &node_image 'node:22-alpine'
-  - &install_deps |
+  - &enable_pnpm 'corepack enable'
-    corepack enable
-    pnpm install --frozen-lockfile
 
 when:
   - event: [push, pull_request, manual]
 
+# Steps run sequentially to avoid OOM on the CI runner.
+# node_modules is installed once by the install step and shared across
+# all subsequent steps via Woodpecker's shared workspace volume.
+
 steps:
   install:
     image: *node_image
     commands:
-      - *install_deps
+      - corepack enable
+      - pnpm install --frozen-lockfile
 
   typecheck:
     image: *node_image
     commands:
-      - *install_deps
+      - *enable_pnpm
       - pnpm typecheck
     depends_on:
       - install
@@ -24,34 +27,31 @@ steps:
   lint:
     image: *node_image
     commands:
-      - *install_deps
+      - *enable_pnpm
       - pnpm lint
     depends_on:
-      - install
+      - typecheck
 
   format:
     image: *node_image
     commands:
-      - *install_deps
+      - *enable_pnpm
       - pnpm format:check
     depends_on:
-      - install
+      - lint
 
   test:
     image: *node_image
     commands:
-      - *install_deps
+      - *enable_pnpm
       - pnpm test
     depends_on:
-      - install
+      - format
 
   build:
     image: *node_image
     commands:
-      - *install_deps
+      - *enable_pnpm
       - pnpm build
     depends_on:
-      - typecheck
-      - lint
-      - format
       - test