RUSTSEC-2026-0462 (GHSA-C9F9-CP8H-JCJ5)
Vulnerability from osv_rustsec – Published: 2026-11-05 12:00 – Updated: 2026-11-09 10:18 – Source websiteBefore 3.9.0, the built-in Streamable HTTP clients in rmcp read non-streaming
HTTP response bodies into memory with no size limit. Both the reqwest client
(transport-streamable-http-client-reqwest) and the Unix socket client
(transport-streamable-http-client-unix-socket) were affected:
application/jsonresponses to POST requests were read in full before parsing.- Non-2xx responses were read in full for error reporting.
A malicious or compromised MCP server can answer with a body that never ends,
for example a chunked application/json response that streams whitespace. The
client buffers it until the process runs out of memory. This works during
initialize, before any authentication. On desktop or agent hosts, this
terminates the whole host application.
Server-sent event streams are not affected. Each event is already capped by
max_sse_event_size, and events are not accumulated.
The flaw was corrected in 3.9.0. Successful JSON bodies are limited to 16 MiB
and non-success bodies to 64 KiB. Bytes are counted while streaming, after
decompression. An oversized Content-Length is rejected before the body is
read, and an oversized body is rejected with
StreamableHttpError::ResponseBodyTooLarge instead of being truncated.
If you cannot upgrade, use a custom StreamableHttpClient that enforces a
response body limit, or connect only to trusted MCP servers.
{
"affected": [
{
"database_specific": {
"categories": [
"denial-of-service"
],
"cvss": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H",
"informational": null
},
"ecosystem_specific": {
"affected_functions": null,
"affects": {
"arch": [],
"functions": [],
"os": []
}
},
"package": {
"ecosystem": "crates.io",
"name": "rmcp",
"purl": "pkg:cargo/rmcp"
},
"ranges": [
{
"events": [
{
"introduced": "0.2.0"
},
{
"fixed": "3.9.0"
}
],
"type": "SEMVER"
}
],
"versions": []
}
],
"aliases": [
"GHSA-c9f9-cp8h-jcj5"
],
"database_specific": {
"license": "CC0-1.0"
},
"details": "Before 3.9.0, the built-in Streamable HTTP clients in rmcp read non-streaming\nHTTP response bodies into memory with no size limit. Both the reqwest client\n(`transport-streamable-http-client-reqwest`) and the Unix socket client\n(`transport-streamable-http-client-unix-socket`) were affected:\n\n- `application/json` responses to POST requests were read in full before\n parsing.\n- Non-2xx responses were read in full for error reporting.\n\nA malicious or compromised MCP server can answer with a body that never ends,\nfor example a chunked `application/json` response that streams whitespace. The\nclient buffers it until the process runs out of memory. This works during\n`initialize`, before any authentication. On desktop or agent hosts, this\nterminates the whole host application.\n\nServer-sent event streams are not affected. Each event is already capped by\n`max_sse_event_size`, and events are not accumulated.\n\nThe flaw was corrected in 3.9.0. Successful JSON bodies are limited to 16 MiB\nand non-success bodies to 64 KiB. Bytes are counted while streaming, after\ndecompression. An oversized `Content-Length` is rejected before the body is\nread, and an oversized body is rejected with\n`StreamableHttpError::ResponseBodyTooLarge` instead of being truncated.\n\nIf you cannot upgrade, use a custom `StreamableHttpClient` that enforces a\nresponse body limit, or connect only to trusted MCP servers.",
"id": "RUSTSEC-2026-0462",
"modified": "2026-11-09T10:18:32Z",
"published": "2026-11-05T12:00:00Z",
"references": [
{
"type": "PACKAGE",
"url": "https://crates.io/crates/rmcp"
},
{
"type": "ADVISORY",
"url": "https://rustsec.org/advisories/RUSTSEC-2026-0462.html"
},
{
"type": "WEB",
"url": "https://github.com/modelcontextprotocol/rust-sdk/commit/3e0156c1eaf1573a1b878e2a8998d2fa2d0b1439"
}
],
"related": [],
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
],
"summary": "Unbounded response body buffering in Streamable HTTP clients"
}
Sightings
| Author | Source | Type | Date | Other |
|---|
Nomenclature
- Seen: The vulnerability was mentioned, discussed, or observed by the user.
- Confirmed: The vulnerability has been validated from an analyst's perspective.
- Published Proof of Concept: A public proof of concept is available for this vulnerability.
- Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
- Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
- Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
- Not confirmed: The user expressed doubt about the validity of the vulnerability.
- Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.
Browse all ATT&CK techniques and the vulnerabilities related to each.
Related by attack behaviour
Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.