Story · Backslash Security Blog
Claude Code's Auto Mode Is Now the Default. It Catches a Lot, But Not Everything. (Backslash Security Blog)
blog post · Story page
Auto Mode caught 89% of dangerous commands in Anthropic's tests and is now the default. Backslash's research is about the rest, and about why an approval mode being automatic is not the same as it being safe.
In plain words
- Claude Code now uses automatic checks by default to decide which computer instructions to allow.
- In Anthropic's tests, those checks caught 89% of dangerous instructions.
- Backslash's research shows that dangerous instructions can still get through, leaving users without a guarantee of safety.
Appeared in
- Agent-written pull requests match human ones on revert rates, and fail differently
Sep 28, 2026 · in the sections
Subscribe
Get the brief in your inbox
Pick daily, weekly, or both. Nothing is gated either way: every issue is on the site and in the feeds.
- Weekdays at 8:45am IST, one lead story and 6 to 9 items.
- Sundays, an argued synthesis rather than a recap.
- One click to leave, and quiet days say so in the subject line.