CVE-2026-34518: 
Wolfi vulnerability analysis and mitigation

CVE-2026-34518 is an information disclosure vulnerability in the aiohttp asynchronous HTTP client/server framework for Python. When following HTTP redirects to a different origin, aiohttp correctly drops the Authorization header but incorrectly retains the Cookie and Proxy-Authorization headers, potentially leaking sensitive credentials to unintended third-party servers. All versions up to and including 3.13.3 are affected; the issue was patched in version 3.13.4. It was disclosed on April 1, 2026, and carries a CVSS v3.1 score of 5.3 (Medium) and a CVSS v4.0 score of 2.7 (Low) (GitHub Advisory, aiohttp Security Advisory).

The root cause is an incomplete header-stripping implementation in aiohttp's redirect-handling logic (aiohttp/client.py), classified as CWE-200 (Exposure of Sensitive Information to an Unauthorized Actor) and CWE-497. When a cross-origin redirect is detected, the code nullified the Authorization header but omitted equivalent handling for Cookie and Proxy-Authorization. The fix (commit 5351c98) adds two additional headers.pop() calls to strip these headers before following the redirect to a different origin. Exploitation requires no authentication, no user interaction, and is reachable over the network from any attacker who can control or influence a redirect target (aiohttp Security Advisory, Patch Commit).

Successful exploitation results in the unintended disclosure of Cookie and Proxy-Authorization header values to a third-party server controlled by an attacker or a malicious redirect target. These headers may contain session tokens, authentication credentials, or proxy credentials, enabling session hijacking or unauthorized access to proxied resources. There is no integrity or availability impact; the vulnerability is limited to confidentiality loss (GitHub Advisory, Red Hat Bugzilla).

There is no public proof-of-concept exploit and no evidence of in-the-wild exploitation at this time. The EPSS score is approximately 0.015% (3rd percentile), indicating a very low probability of near-term exploitation. The vulnerability is not listed in the CISA Known Exploited Vulnerabilities (KEV) catalog. No threat actor attribution has been reported (GitHub Advisory).

1. Set up a malicious redirect server: The attacker controls or compromises a server (origin A) that issues an HTTP redirect (e.g., 301/302) pointing to a different origin (origin B) also under attacker control.
2. Induce the victim application to make a request: The attacker causes a vulnerable aiohttp client application to send an HTTP request to origin A, including Cookie and/or Proxy-Authorization headers (e.g., as part of normal authenticated API usage).
3. Trigger the cross-origin redirect: Origin A responds with a redirect to origin B (a different host/scheme). The vulnerable aiohttp client follows the redirect automatically.
4. Capture leaked headers: Because aiohttp fails to strip Cookie and Proxy-Authorization before following the redirect, origin B receives these headers in the forwarded request, allowing the attacker to harvest session tokens or proxy credentials.
5. Leverage captured credentials: The attacker uses the stolen session cookie or proxy credentials to impersonate the victim or access proxied resources (aiohttp Security Advisory, Patch Commit).

• Network: Outbound HTTP requests from an aiohttp client to an unexpected or attacker-controlled domain following a redirect, with Cookie or Proxy-Authorization headers present in the forwarded request.
• Logs: Application or proxy logs showing cross-origin redirect chains where sensitive headers are forwarded to a different host than the original request target.
• Logs: Access logs on third-party servers receiving Cookie or Proxy-Authorization headers that were not explicitly sent by the application to that origin.

Upgrade aiohttp to version 3.13.4 or later, which drops Cookie and Proxy-Authorization headers on cross-origin redirects (commit 5351c98). No configuration-based workaround is available in older versions; the only remediation is upgrading. IBM products incorporating aiohttp (Process Mining, API Connect, watsonx Code Assistant On Prem, Instana Observability) have also released interim fixes addressing this CVE (aiohttp Release, IBM Process Mining, IBM API Connect).

The vulnerability was reported by researcher uug4na and remediated by Dreamsorcerer from the aiohttp maintainer team, who published the advisory and patch on April 1, 2026. Red Hat opened a tracking bug (Bugzilla #2454098) and classified the severity as Low. IBM issued security bulletins for multiple affected products including Process Mining, API Connect, watsonx Code Assistant, and Instana Observability. No significant broader media coverage or notable community controversy has been observed (aiohttp Security Advisory, Red Hat Bugzilla, IBM Process Mining).