fix(rebuild): harden split workflow setup (#164)

## Motivation

Weekly rebuild shards can all hit a transient CRAN DNS/index outage at once, and the dependent CDN purge always fails because it tries to clone over the checkout preserved from the re-index step.

## Changes

- Retry `uvr add` resolution up to four times with bounded backoff.
- Reuse the existing Crow workspace checkout in the CDN purge step.
- Remove the purge step's unused Git package and repository token.

## Validation

- `crow lint .crow/`
- `shellcheck local/uvr-install.sh scripts/purge_cdn_zone.sh`
- `git diff --check`

Reviewed-on: #164

local/uvr-install.sh

@@ -84,9 +84,20 @@ trap 'rm -rf "$project_dir"' EXIT
 cd "$project_dir"
 
 "$uvr_bin" init --here --r-version "$r_full"
-# --no-install: resolve and lock only. The install happens in the sync below,
-# which is the only command that honours --library.
-"$uvr_bin" add --no-install "$@"
+# --no-install resolves and locks only; retry because concurrent shards can
+# expose short-lived DNS or CRAN-index failures and uvr rolls the manifest back
+# cleanly after an unsuccessful resolution.
+add_attempt=1
+while ! "$uvr_bin" add --no-install "$@"; do
+  if [ "$add_attempt" -ge 4 ]; then
+    echo "error: uvr add failed after ${add_attempt} attempts" >&2
+    exit 1
+  fi
+  add_delay=$((add_attempt * 10))
+  echo "warning: uvr add attempt ${add_attempt} failed; retrying in ${add_delay}s" >&2
+  sleep "$add_delay"
+  add_attempt=$((add_attempt + 1))
+done
 
 # TEMPORARY (drop once the images ship a uvr above v0.4.5): the sync below runs
 # `apt-get install` for every resolved system dependency without refreshing the