Unclear It is not clear whether the patterns themselves changed, so whether these rules now flag different text is unknown.
What
Claude Code has a secret scanner, a check that looks through text for things such as passwords, keys or tokens. Two of its rules are marked low-confidence, meaning a match is a weaker sign of a real secret. They now load ready-made search patterns instead of building their patterns in place:
- the rule for sensitive-looking values assigned to a name (
sensitive-assign) - the rule for cloud environment variables
Why
This looks like a tidy-up of how the patterns are stored. If the patterns themselves differ, what these two rules flag may differ too.
Something disagreesSomething we can check disagrees with this entry, or the writer said they could not settle it.
The writer flagged doubt
It is not clear whether the patterns themselves changed, so whether these rules now flag different text is unknown.