@@ -108,7 +132,16 @@ const V2CatchUpStrip: React.FC
= ({ podId }) => {
>
{refreshing
? t('podChat.catchup.working')
- : summary ? t('podChat.catchup.refresh') : t('podChat.catchup.summarize')}
+ : t('podChat.catchup.refresh')}
+
+
{expanded && summary?.content && (
diff --git a/frontend/src/v2/v2.css b/frontend/src/v2/v2.css
index 9a5376df5..ce1df52f1 100644
--- a/frontend/src/v2/v2.css
+++ b/frontend/src/v2/v2.css
@@ -1715,6 +1715,16 @@ body.modern-ui.v2-canvas {
}
.v2-root button.v2-catchup__refresh:hover:not(:disabled) { text-decoration: underline; }
.v2-root button.v2-catchup__refresh:disabled { color: var(--v2-text-tertiary); cursor: default; }
+.v2-root button.v2-catchup__dismiss {
+ background: transparent;
+ border: none;
+ padding: 0 2px;
+ font-size: 14px;
+ line-height: 1;
+ color: var(--v2-text-tertiary);
+ cursor: pointer;
+}
+.v2-root button.v2-catchup__dismiss:hover { color: var(--v2-text); }
.v2-catchup__body {
margin-top: 6px;
max-height: 220px;
diff --git a/k8s/helm/commonly/templates/core/backend-deployment.yaml b/k8s/helm/commonly/templates/core/backend-deployment.yaml
index 94f5b3eea..14897a220 100644
--- a/k8s/helm/commonly/templates/core/backend-deployment.yaml
+++ b/k8s/helm/commonly/templates/core/backend-deployment.yaml
@@ -419,7 +419,10 @@ spec:
- name: LITELLM_BASE_URL
value: "http://litellm:4000"
- name: LITELLM_CHAT_MODEL
- value: "chatgpt/gpt-5.4-mini"
+ # deepseek-v4-flash (Sam, 2026-09-01): summary quality on
+ # gpt-5.4-mini was "no good info at all"; the DeepSeek flash
+ # entry already exists in LiteLLM with a live key.
+ value: "deepseek-v4-flash"
# Temporary: bypass LiteLLM's broken chatgpt/ bridge (BerriAI/litellm#25429)
# and route Codex directly via OpenClaw's native openai-codex-responses handler.
# Flip to "false" / remove once LiteLLM ships a fix for empty-output responses.