Cluster rebuilt on Substrate v0.5.0 / AX 7b3347f: new CLI, one workspace per task, digest-pinned images

README.md

@@ -6,6 +6,11 @@ Qwen 3.8 27B model that Alex hosts. No API keys or accounts needed. The
 recommended agent is [pi](https://pi.dev); see [Using pi with
 Qwen](#5-using-pi-with-qwen). Claude Code is available too.
 
+> **Used this before 2026-10-21?** The cluster was rebuilt on new versions, and
+> old tasks and their workspaces are gone. Install the new CLI (step 2). In
+> your `task.yaml`, `workspaces:` (a list) is now `workspace:` (one entry), and
+> `image:` must be pinned by digest. Copy the example below.
+
 ## 1. Join
 
 1. Install [Tailscale](https://tailscale.com/download) and sign in with any
@@ -21,7 +26,7 @@ Pick your platform: `darwin-arm64` (Apple Silicon), `darwin-amd64`,
 
 ```bash
 P=darwin-arm64
-B=https://code.proc.io/api/packages/ax/generic/ax-cli/ac23328-proc.3
+B=https://code.proc.io/api/packages/ax/generic/ax-cli/7b3347f-proc.1
 curl -fLO $B/ax-$P && curl -fLO $B/SHA256SUMS
 shasum -a 256 -c SHA256SUMS --ignore-missing
 chmod +x ax-$P && sudo mv ax-$P /usr/local/bin/ax
@@ -66,11 +71,11 @@ metadata:
   name: sam-hello
   atespace: default
 spec:
-  # pi:          code.proc.io/ax/ax-task-runner:pi-vllm-3
+  # pi:          code.proc.io/ax/ax-task-runner:pi-vllm-4@sha256:7fd48625ba206e298c46a292d0c1197c7b5e63f756fbc0b65d23b09e30d5c82a
-  # Claude Code: code.proc.io/ax/ax-task-runner:claude-vllm-3
+  # Claude Code: code.proc.io/ax/ax-task-runner:claude-vllm-4@sha256:1fb4d8d08ef6b25e1bb879260f049ba612ba599a49c7b24a24990fb8608a96a1
-  image: code.proc.io/ax/ax-task-runner:pi-vllm-3
+  image: code.proc.io/ax/ax-task-runner:pi-vllm-4@sha256:7fd48625ba206e298c46a292d0c1197c7b5e63f756fbc0b65d23b09e30d5c82a
-  workspaces:
+  workspace:
-    - name: sam-myrepo
+    name: sam-myrepo
     path: /workspace
   debug: true          # required for `ax ssh`
 ```
@@ -131,7 +136,7 @@ When the server is busy (or for unrelated questions), start fresh with
 
 ### pi is preconfigured
 
-The pi image (`ax-task-runner:pi-vllm-3`) points pi at that model, so there's
+The pi image (`ax-task-runner:pi-vllm-4`) points pi at that model, so there's
 nothing to log in to or set up. pi runs in the task's working directory with
 four tools: `read`, `bash`, `edit` and `write`. It can run builds and tests and
 change files in the repo.
@@ -190,7 +195,7 @@ nearest supported level.
 
 ### Claude Code instead
 
-Set the image to `code.proc.io/ax/ax-task-runner:claude-vllm-3` and use
+Set the image to `code.proc.io/ax/ax-task-runner:claude-vllm-4@sha256:1fb4d8d08ef6b25e1bb879260f049ba612ba599a49c7b24a24990fb8608a96a1` and use
 `claude -p "…"` (also against Qwen). In our tests, pi was faster on the same
 model. Claude Code's `--continue` history doesn't survive a suspend; pi's does.
 

task.yaml

@@ -17,10 +17,10 @@ metadata:
   name: sam-hello
   atespace: default
 spec:
-  # pi:          code.proc.io/ax/ax-task-runner:pi-vllm-3
+  # pi:          code.proc.io/ax/ax-task-runner:pi-vllm-4@sha256:7fd48625ba206e298c46a292d0c1197c7b5e63f756fbc0b65d23b09e30d5c82a
-  # Claude Code: code.proc.io/ax/ax-task-runner:claude-vllm-3
+  # Claude Code: code.proc.io/ax/ax-task-runner:claude-vllm-4@sha256:1fb4d8d08ef6b25e1bb879260f049ba612ba599a49c7b24a24990fb8608a96a1
-  image: code.proc.io/ax/ax-task-runner:pi-vllm-3
+  image: code.proc.io/ax/ax-task-runner:pi-vllm-4@sha256:7fd48625ba206e298c46a292d0c1197c7b5e63f756fbc0b65d23b09e30d5c82a
-  workspaces:
+  workspace:
-    - name: sam-myrepo
+    name: sam-myrepo
     path: /workspace
   debug: true          # required for `ax ssh`