CVE-2021-28363

A vulnerability in urllib3 causes the Proxy-Authorization header not to be stripped on cross-origin redirects when not using urllib3's built-in proxy support. This could leak credentials to a malicious server under specific conditions. The severity is low and requires misconfiguration.

urllib3 versions from 1.10.3 until 2.8.0 contain an unbounded memory allocation vulnerability. The HTTPResponse.read_chunked and HTTPResponse.stream methods buffer the chunk-size field without a length bound until a newline or EOF is encountered. A malicious HTTP server can send a very long unterminated chunk-size line, leading to exhaustion of the client process memory.

In urllib3 (versions 2.6.2 through 2.8.0), the HTTPResponse.stream and HTTPResponse.read_chunked methods can enter an infinite loop because the Deflate decoder retains trailing bytes as unconsumed input after reaching end-of-stream and repeatedly decodes them without progress. The issue occurs when an untrusted server sends a chunked Deflate response whose decoded body exceeds a positive finite chunk size and whose encoded body has trailing bytes.

urllib3 versions from 1.26.0 until 2.8.0 fail to keep target-server TLS settings separate from the HTTPS proxy connection. The proxy_ssl_context, proxy_assert_hostname, proxy_assert_fingerprint, ssl_context, cert_reqs, verify_mode, use_forwarding_for_https=True, and CERT_NONE configuration paths are affected. Setting cert_reqs=CERT_NONE can overwrite proxy_ssl_context.verify_mode in place, and the mutation persists so later connections may connect to the HTTPS proxy without certificate verification.

The urllib3 HTTP client library for Python versions 2.6.0 to before 2.7.0 contains a vulnerability where the entire response is decompressed instead of the requested portion during the second HTTPResponse.read(amt=N) call when using the official Brotli library, or after HTTPResponse.drain_conn() is called after partial read. This can lead to excessive resource consumption (high CPU usage and memory allocation) on the client side.

The urllib3 Python library versions 1.22 through 2.6.2 incorrectly decompresses entire HTTP redirect response bodies even when using the streaming API. This enables a decompression bomb attack leading to excessive client resource consumption.

urllib3 is a user-friendly HTTP client library for Python. Starting in version 1.0 and prior to 2.6.0, the Streaming API improperly handles highly compressed data. When streaming a compressed response, urllib3 can perform decoding or decompression based on the HTTP Content-Encoding header (e.g., gzip, deflate, br, or zstd). The decompression logic could cause urllib3 to fully decode a small amount of highly compressed data in a single operation, resulting in excessive resource consumption (high CPU usage and massive memory allocation).

urllib3 is a user-friendly HTTP client library for Python. Starting in version 1.24 and prior to 2.6.0, the number of links in the decompression chain was unbounded allowing a malicious server to insert a virtually unlimited number of compression steps leading to high CPU usage and massive memory allocation for the decompressed data. This vulnerability is fixed in 2.6.0.