Error-retry path appears to rebuild the prefix
PrefixAudit rule retry-rebuilds · v0.1.0
Why it breaks your cache
A structured-output retry loop that resends a fresh system prompt pays a new write every time. A 98.4% structured-output success rate can still hide a 2% retry loop eating 12-18% of the inference budget.
How to fix it
Append the parse error and a corrective instruction as a new user message after the cached span. Never rebuild and resend the system prompt to retry.
How common is it?
This rule needs request or runtime context that an extracted prompt cannot supply, so it is excluded from the corpus study. It still runs in the auditor when you supply that context.
Estimated impact
Modelled as invalidating 0% of the cached prefix.
Is your prompt affected?
Paste it and find out. Static analysis, in-browser, nothing leaves the page.
Run the audit →