chore(sync): mirror docs from openclaw/openclaw@fca1c6e

Author: openclaw-docs-sync[bot]

.openclaw-sync/source.json

@@ -1,15 +1,15 @@
 {
   "repository": "openclaw/openclaw",
-  "sha": "8bba2cf517a2e971ff8550ff7f4053b8e5f2bccf",
+  "sha": "fca1c6ec13f40534e288b1e02eee472cbbdfc34d",
   "sources": {
     "openclaw": {
       "repository": "openclaw/openclaw",
-      "sha": "8bba2cf517a2e971ff8550ff7f4053b8e5f2bccf"
+      "sha": "fca1c6ec13f40534e288b1e02eee472cbbdfc34d"
     },
     "clawhub": {
       "repository": "openclaw/clawhub",
-      "sha": "ad35a92c5af36d4e97004c40b70ba4e273013a16"
+      "sha": "22252375c61a07466e470d54f30b18b666b2c4e4"
     }
   },
-  "syncedAt": "2026-11-02T05:29:05.205Z"
+  "syncedAt": "2026-11-02T06:00:41.545Z"
 }

docs/clawhub/cli.md

@@ -310,7 +310,7 @@ clawhub scan download @scope/demo --version 2.0.0 --kind plugin --output report.
 #### GitHub Actions
 
 ClawHub ships an official reusable workflow at
-[`/.github/workflows/skill-publish.yml`](https://github.com/openclaw/clawhub/blob/ad35a92c5af36d4e97004c40b70ba4e273013a16/.github/workflows/skill-publish.yml)
+[`/.github/workflows/skill-publish.yml`](https://github.com/openclaw/clawhub/blob/22252375c61a07466e470d54f30b18b666b2c4e4/.github/workflows/skill-publish.yml)
 for skill repos and catalog repos.
 
 Typical catalog setup:
@@ -844,7 +844,7 @@ Notes:
 #### GitHub Actions
 
 ClawHub also ships an official reusable workflow at
-[`/.github/workflows/package-publish.yml`](https://github.com/openclaw/clawhub/blob/ad35a92c5af36d4e97004c40b70ba4e273013a16/.github/workflows/package-publish.yml)
+[`/.github/workflows/package-publish.yml`](https://github.com/openclaw/clawhub/blob/22252375c61a07466e470d54f30b18b666b2c4e4/.github/workflows/package-publish.yml)
 for plugin repos.
 
 Typical caller setup:

docs/concepts/usage-tracking.md

@@ -104,9 +104,10 @@ order of floating-point cost additions. Cached SQLite detail reads use the rollu
 recorded checkpoint and current transcript mutation facts instead of rescanning
 all transcript rows for their size and event count.
 
-Background refresh budgets use elapsed time, so wall-clock corrections do not
-extend the quiet period or retry delays. Report dates and recorded timestamps
-still follow the wall clock.
+Background refresh budgets, detailed-result cache freshness, and the Control UI's
+context-breakdown retry backoff use elapsed time, so wall-clock corrections do not
+extend the quiet period, keep expired results fresh, or shorten retry delays.
+Report dates and recorded timestamps still follow the wall clock.
 
 ## Usage date ranges