CVE-2026-16756: Slowloris Denial of Service via Resource Exhaustion in aws-smithy-http-server
Vulnerability ID: CVE-2026-16756
CVSS Score: 7.5
Published: 2026-07-24
An unauthenticated remote resource exhaustion vulnerability in Amazon aws-smithy-http-server enables denial-of-service (DoS) attacks. Affected versions do not enforce connection limits or header timeouts, allowing standard Slowloris techniques to block server operations.
TL;DR
Unauthenticated attackers can exhaust server connections and task buffers in Amazon aws-smithy-http-server by maintaining slow, partial HTTP request streams, leading to a complete denial-of-service condition.
⚠️ Exploit Status: POC
Technical Details
- CWE ID: CWE-770 (Allocation of Resources Without Limits or Throttling)
- Attack Vector: Network
- CVSS: 7.5 (High)
- EPSS Score: 0.00417 (34.15th Percentile)
- Impact: Denial of Service (DoS) via resource exhaustion
- Exploit Status: Proof-of-Concept / Theoretical
- KEV Status: Not Listed
Affected Systems
- AWS Smithy Runtime Systems
- Rust Microservices utilizing smithy-rs server generation
- Deployments running stand-alone aws-smithy-http-server binaries
-
aws-smithy-http-server: <= 0.66.4 (Fixed in:
0.66.5)
Mitigation Strategies
- Upgrade the aws-smithy-http-server library dependency to version 0.66.5 or later.
- Place a dedicated reverse proxy (NGINX, HAProxy) or Application Load Balancer in front of the application.
- Apply localized connection limits and header timeouts using custom Tower or Tokio middleware stack configurations.
Remediation Steps:
- Open Cargo.toml within the Rust workspace containing the Smithy server logic.
- Update the dependency version of
aws-smithy-http-serverto"0.66.5"or newer. - Run cargo update to download the patched crate version and rebuild the application binary.
- Validate the fix locally by setting a low connection timeout and observing server response to incomplete headers.
References
Read the full report for CVE-2026-16756 on our website for more details including interactive diagrams and full exploit analysis.
Top comments (0)