{"version":"2.1.285","anchor":"secret-redaction-hardened-around-truncation-boundaries","canonical_anchor":"secret-redaction-hardened-around-truncation-boundaries","heading":"Shortened output no longer cuts a hidden secret in half","tier":"notice","area":"Redaction","url":"https:\/\/changelogs.core-directive.com\/v\/2.1.285\/e\/secret-redaction-hardened-around-truncation-boundaries","release_url":"https:\/\/changelogs.core-directive.com\/v\/2.1.285","markdown":"### Shortened output no longer cuts a hidden secret in half\n\nWhen Claude Code shortens output and hides secrets, it now moves its cut points so it does not split a secret and leave part of it showing\n\n**Unclear.** Which output this applies to is not settled.\n\n**What**\n\nClaude Code can shorten long output and replace secrets such as passwords or tokens with `[REDACTED]`. When it shortens text, it now moves the cut so it does not land in the middle of a secret. It watches for the signs that often come before one, such as `=`, `:`, a quote mark, `bearer` or `basic`. It also keeps count of how many characters were removed. Before, it simply cut at a line break and kept the end.\n\n**Why**\n\nA shortened log or output tail is less likely to show part of a credential.\n\n- Area: Redaction\n- Tier: You'll notice\n- Useful: 2\/5\n- Signal: 1\/5"}