Starlet web server for Perl sends unwanted data on HEAD requests
A small web server for Perl can send extra data in replies to some requests. This can mix up responses on connections that stay open, so a visitor could receive another request's reply.
| Severity | HighVulnerability |
|---|---|
| Being exploited? | UnclearNo reports either way yet |
| Fix available? | YesStarlet 0.33 or later · since 30 Oct 2026 |
| How sure is this? | Official, 2 sourcesBacked by an official source |
How the attack works simplified
- 1Client sends HEADRequest for headers only
- 2Starlet repliesSends an unexpected body
- 3Connection reusedLeftover bytes stay on the wire
- 4Next response mixedReader takes extra bytes as next reply
- ResultResponses on shared connections get desynchronized
What happened
Starlet versions before 0.33 for Perl send a response body for HEAD requests, and do not pass the request method to the response writer. The extra bytes corrupt the next response on persistent or pipelined connections.
Are you affected?
You are affected if you run Starlet for Perl versions before 0.33 serving HTTP/1.1 applications with persistent or pipelined connections, especially behind a proxy that shares connections.
Starlet for Perl, all versions before 0.33
What to do
- Update Starlet for Perl to version 0.33 or later
- If you cannot update right away, put a proxy in front that does not reuse connections across clients
- Check that your applications and middleware handle HEAD requests correctly
Timeline
- 30 Oct 2026, 13:18 UTCPublicly disclosed
- 30 Oct 2026, 15:45 UTCReported by oss-security · oss-security
- 30 Oct 2026, 16:00 UTCPicked up by Zero Radar
- 30 Oct 2026, 17:22 UTCNVD reported on it · NVD
Sources 2 reports, oldest first
Official vendor, government CERT or advisory database
- CVE-2026-116127NVD · 30 Oct 2026, 13:18 UTC
Security researchers research teams and mailing lists
- Starlet versions before 0.33 for Perl (4 CVEs)oss-security · 30 Oct 2026, 15:45 UTC
Technical details
- ID
- CVE-2026-116127
- Disclosed
- 30 Oct 2026, 13:18 UTC
- On Zero Radar
- 30 Oct 2026, 16:00 UTC (3 h after disclosure)
- Last change
- 30 Oct 2026, 17:22 UTC
The plain-language parts and the attack diagram were written by an AI model (anthropic/claude-haiku-5.5) from the sources above. Facts like versions, exploitation and dates come from the sources themselves. How Zero Radar works