Skip to main content
Vulnerability Database/CVE-2026-88382

CVE-2026-88382: Hiredis RESP Buffer Overflow Vulnerability

CVE-2026-88382 is a buffer overflow flaw in Hiredis RESP aggregate parser that enables uncontrolled memory allocation. This article covers the technical details, affected versions, security impact, and mitigation strategies.

Published:

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

CVE-2026-88382 Overview

CVE-2026-88382 is an uncontrolled memory allocation vulnerability in hiredis, the minimalistic C client library for the Redis database. The flaw exists in the RESP (REdis Serialization Protocol) aggregate parser in commit 29ea279, a post-v1.5.0 build. A remote attacker can send a crafted RESP aggregate reply that triggers unbounded memory allocation on the client side. The condition leads to resource exhaustion and process termination, resulting in denial of service. The weakness is classified under CWE-770: Allocation of Resources Without Limits or Throttling.

Critical Impact

Remote attackers can exhaust memory in any application that uses the affected hiredis build to connect to an attacker-controlled or compromised Redis endpoint, causing availability loss.

Affected Products

  • hiredis commit 29ea279 (post-v1.5.0)
  • Applications and services linking the affected hiredis build
  • Downstream Redis client tooling embedding the vulnerable parser

Discovery Timeline

  • 2026-09-24 - CVE-2026-88382 published to the National Vulnerability Database
  • 2026-09-24 - Last updated in NVD database

Technical Details for CVE-2026-88382

Vulnerability Analysis

The vulnerability resides in the hiredis RESP aggregate parser, which handles multi-bulk and array replies from a Redis server. RESP aggregate types carry a length prefix that indicates the number of elements the client should expect. The parser uses this attacker-controlled length to allocate memory for the reply structure without validating it against the number of bytes actually available on the wire. A malicious server can advertise an aggregate reply with an extremely large element count while sending minimal data. The client then attempts to allocate memory sized by the declared count, which exhausts the process heap and terminates the client.

Root Cause

The root cause is missing bounds enforcement on untrusted length fields parsed from RESP replies [CWE-770]. The parser trusts the declared aggregate size and issues an allocation proportional to that value before verifying payload consistency. No upper bound, quota, or sanity ratio between declared length and received bytes is applied.

Attack Vector

Exploitation requires network reachability between the vulnerable client and an attacker-controlled Redis endpoint. The attacker can host a rogue Redis server, hijack a legitimate server, or inject replies via a man-in-the-middle position on an unencrypted connection. Any hiredis-based application that connects to the hostile endpoint will parse the malicious aggregate reply and crash. No authentication or user interaction on the client side is required. Full technical discussion is tracked in the upstream GitHub Hiredis Issue #1357.

// No verified public exploit code is available.
// See the upstream issue tracker for reproducer details.

Detection Methods for CVE-2026-88382

Indicators of Compromise

  • Unexpected termination or SIGABRT/out-of-memory kills in processes linked against hiredis
  • Kernel OOM-killer events coinciding with outbound connections to Redis TCP port 6379
  • Redis client connections to untrusted or newly observed external IP addresses
  • RESP replies containing abnormally large aggregate length prefixes relative to payload size

Detection Strategies

  • Inventory binaries and container images to identify those statically or dynamically linked against hiredis commit 29ea279.
  • Inspect RESP traffic at network choke points for aggregate headers (*, %, ~, >) declaring element counts that exceed reasonable application thresholds.
  • Correlate repeated client process crashes with recent Redis connection activity in host telemetry.

Monitoring Recommendations

  • Alert on OOM-killer and segmentation fault events for services known to use hiredis.
  • Monitor outbound connections from application tiers to Redis endpoints outside the approved allowlist.
  • Track memory consumption spikes in Redis client processes for rapid growth preceding termination.

How to Mitigate CVE-2026-88382

Immediate Actions Required

  • Identify all applications linking the affected hiredis commit 29ea279 and prioritize them for patching.
  • Restrict Redis client connections to trusted, authenticated endpoints only.
  • Enforce TLS for all Redis traffic to prevent man-in-the-middle injection of malicious RESP replies.

Patch Information

At the time of publication, the NVD entry references the upstream GitHub Hiredis Issue #1357 for remediation status. Rebuild and redeploy any dependent software against a fixed upstream commit once published by the hiredis maintainers.

Workarounds

  • Limit the memory available to processes using hiredis with ulimit -v or cgroup memory limits so a crash does not destabilize the host.
  • Place a Redis-aware proxy between clients and servers to reject replies with implausible aggregate length prefixes.
  • Isolate Redis infrastructure on private networks and enforce firewall rules preventing clients from reaching untrusted servers.
bash
# Example cgroup v2 memory cap for a hiredis-based service
sudo systemctl set-property my-redis-client.service MemoryMax=512M
sudo systemctl daemon-reload
sudo systemctl restart my-redis-client.service

Disclaimer: This content was generated using AI. While we strive for accuracy, please verify critical information with official sources.

Sources: National Vulnerability Database, CISA KEV, Exploit-DB

Experience the Most Advanced Cybersecurity Platform

See how the world’s most intelligent, autonomous cybersecurity platform can protect your organization today and into the future.