cowlib

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Improper Encoding or Escaping of Output vulnerability in ninenines cowlib allows Link header directive smuggling via unescaped special characters in cow_link:link/1.

cow_link:do_link/1 in cowlib interpolates the target URI, rel value, and attribute keys directly into the serialized Link: header value without escaping or token-grammar validation. A > byte in target prematurely closes the URI slot, allowing an attacker to append additional link entries with attacker-chosen rel directives. A " or \ in rel escapes the quoted string and opens new parameters. Any byte — including whitespace, =, and " — in an attribute key is emitted verbatim. Because browsers act on Link: directives such as rel="preconnect", rel="preload", and rel="prerender", an attacker who can influence these fields in an application that round-trips parsed Link headers through cow_link:link/1 can force victim browsers to make out-of-band connections to attacker-controlled origins.

This issue affects cowlib: from 2.9.0 onward.

Workarounds

Sanitize values before passing them to cow_link:link/1: reject or strip any target value containing >, any rel value containing " or \, and any attribute key containing characters outside the HTTP token grammar. Alternatively, ensure all Link entry fields are derived exclusively from trusted, application-controlled values rather than untrusted input.

Configurations

The application must pass values that originate from, or can be influenced by, untrusted input into cow_link:link/1 — specifically the target URI, rel value, or attribute keys. Applications that construct Link entries exclusively from trusted, application-controlled values are not affected.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Summary

Improper Neutralization of CRLF Sequences in HTTP Headers ('HTTP Request/Response Splitting') vulnerability in ninenines cowlib allows HTTP response splitting via non-VCHAR bytes in structured-fields string values.

cow_http_struct_hd:escape_string/2 in cowlib only escapes \ and ", passing all other bytes through verbatim. This creates an encoder/decoder asymmetry: the matching parser accepts only printable ASCII (0x20–0x7E, excluding " and \), but the encoder emits any byte including CR and LF. An application that builds a structured HTTP header via cow_http_struct_hd:item/1 (or a higher-level wrapper such as cow_http_hd:wt_protocol/1) from attacker-controlled input can have \r\n injected into the serialized header value. Once on the wire, the injected CRLF terminates the current header and any following bytes are interpreted as a new header, enabling HTTP response splitting.

This issue affects cowlib from 2.9.0.

Workarounds

Validate all values passed into structured-fields header builders (directly via cow_http_struct_hd:item/1 or indirectly via higher-level wrappers) before calling the encoder. Reject any value that is not from a trusted, application-controlled source or that contains CR (\r) or LF (\n) bytes.

Applications using cowboy 2.16.0 or later are protected on the server side by the invalid_response_headers option (defaults to error_terminate), which rejects any outgoing response header value containing CR or LF before it reaches the wire. Applications using gun 2.4.0 or later are protected on the client side by the invalid_request_headers request option (defaults to raise), which raises an exception when an outgoing request header value contains CR or LF.

Configurations

The application must pass attacker-controlled data as a string value into cow_http_struct_hd:item/1 (or a wrapper that delegates to it). Applications that construct structured-fields header values exclusively from trusted, application-controlled values are not affected.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows HTTP request splitting and cookie smuggling via unvalidated cookie name and value fields.

cow_cookie:cookie/1 in cowlib builds a client-side Cookie: request header from a list of name-value pairs without validating either field. An attacker who controls the cookie names or values passed to this function can inject ;, ,, CR, LF, or TAB characters into the serialized header. This enables two classes of attack: cookie smuggling within a single header (e.g. injecting "; admin=1" to introduce a phantom cookie that the receiving server treats as authentic) and HTTP request header splitting (injecting CRLF to append arbitrary headers or smuggle a complete second request against a shared upstream proxy). The decoder side (parse_cookie_name/1, parse_cookie_value/1) and setcookie/3 already validate and reject these characters; the encoder alone is missing the check.

This issue affects cowlib from 2.9.0.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.

Improper Handling of Highly Compressed Data (Data Amplification) vulnerability in ninenines cowlib allows unauthenticated remote denial of service via memory exhaustion.

cow_spdy:inflate/2 in cowlib passes peer-supplied compressed bytes directly to zlib:inflate/2 with no output size bound. The SPDY header compression dictionary (?ZDICT) is public, and zlib compresses long runs of repeated bytes at roughly 1024:1, so a few kilobytes of SPDY frame payload can decompress to gigabytes on the BEAM heap, OOM-killing the node. A single unauthenticated SPDY frame is sufficient to trigger the condition. The parsers for syn_stream, syn_reply, and headers frame types are all affected via cow_spdy:parse_headers/2.

This issue affects cowlib from 0.1.0 before 2.16.1.

Uncontrolled Resource Consumption vulnerability in ninenines cowlib (cow_http_te module) allows Excessive Allocation.

The chunked transfer-encoding parser in cow_http_te accepts an unbounded number of hex digits in the chunk-size field. Each digit causes a bignum multiplication (Len * 16 + digit), so parsing N hex digits requires O(N²) CPU work and O(N) memory. Additionally, when input is drip-fed, the parser discards the accumulated length on each partial read and restarts from zero on resumption, raising the cost to O(N³). An unauthenticated remote attacker can exploit this by sending an HTTP/1.1 request with Transfer-Encoding: chunked and a very long chunk-size hex string to cause denial of service through CPU exhaustion and memory amplification.

This vulnerability is associated with program file src/cow_http_te.erl and program routines cow_http_te:stream_chunked/2, cow_http_te:chunked_len/4.

This issue affects cowlib: from 0.6.0 before 2.16.1.

Improper Neutralization of CRLF Sequences ('CRLF Injection') vulnerability in ninenines cowlib allows SSE event splitting and injection via unvalidated field values.

cow_sse:event/1 in cowlib guards the id and event fields against \n but not against bare \r, and the internal prefix_lines/2 function used for data and comment fields splits only on \n. Because the SSE specification requires decoders to treat \r\n, \r, and \n as equivalent line terminators, an attacker who controls any of these fields can inject additional SSE lines and forge a complete event with an arbitrary event type and data payload on the receiving end. In typical deployments where browser EventSource clients or other SSE consumers dispatch on event.type and render event.data, this enables event splitting, client-side logic manipulation, and stored-XSS-equivalent behaviour when event data is inserted into the DOM.

This issue affects cowlib from 2.6.0.

Summary

Allocation of resources without limits vulnerability in ninenines cowlib allows an unauthenticated remote HTTP/2 or HTTP/3 peer to exhaust memory on the vulnerable server (or client) and cause a denial of service.

The HPACK and QPACK prefixed-integer decoder cow_hpack_common:dec_big_int/3 in src/cow_hpack_common.hrl (invoked from cow_hpack:decode/2 in src/cow_hpack.erl and from cow_qpack:decode_field_section/3 in src/cow_qpack.erl) reads continuation octets until it sees one whose high bit is clear, evaluating Int + (Value bsl M) at each step with the shift M growing by seven per octet. No limit is enforced on the number of continuation octets, on the resulting bit width, or on the value; the decoder consumes whatever encoded length the peer supplies.

Because Erlang integers are immutable, each intermediate Value bsl M and each accumulator update allocates a fresh bignum whose digit width grows linearly with the number of octets processed so far. Summed across the whole decode, the transient bignum digit materialization is on the order of the square of the encoded length. A single maximal HPACK indexed representation carried inside one HTTP/2 HEADERS plus one CONTINUATION frame at Cowboy's default max_frame_size_received can force hundreds of megabytes of transient allocation and garbage-collection churn before the resulting header-table index is rejected as invalid. Repeated or concurrent connections multiply the pressure and can drive the Erlang VM to memory exhaustion.

Cowlib is the HTTP parser used by Cowboy, RabbitMQ's management plugin, and other Erlang and Elixir HTTP/2 and HTTP/3 servers and clients, so any exposed endpoint that accepts HPACK or QPACK from an untrusted peer is reachable.

This issue affects cowlib: from 2.0.0 before 2.19.0.