GHSA-RW4J-R22C-9GC3
Vulnerability from github – Published: 2026-09-17 14:53 – Updated: 2026-09-17 14:53Summary
A malicious SSH server can wedge an AsyncSSH client, and an authenticated
client can wedge an AsyncSSH server, by sending a channel maximum packet
size of 0 in SSH_MSG_CHANNEL_OPEN_CONFIRMATION (server→client) or
SSH_MSG_CHANNEL_OPEN (client→server). AsyncSSH stores the peer-supplied value
verbatim with no lower-bound check; the first time channel data is written,
SSHChannel._flush_send_buf enters a synchronous infinite loop that cannot
be interrupted by asyncio.wait_for or any timeout. The loop body has no
await, so it blocks the entire asyncio event loop — for a server, one
malicious authenticated channel freezes all current and future connections.
RFC 4254 §5.1 leaves receiver behavior for a peer-reported "maximum packet size = 0" undefined, so the value must be rejected rather than stored.
Root cause
asyncssh/channel.py:
# process_open (server side) -- line 465
self._send_pktsize = send_pktsize # peer value, no >= 1 check
# process_open_confirmation (client) -- line 528
self._send_pktsize = send_pktsize # peer value, no >= 1 check
# _flush_send_buf -- lines 305-320
while self._send_buf and self._send_window:
pktsize = min(self._send_window, self._send_pktsize) # 0 when peer sends 0
buf, datatype = self._send_buf[0]
if len(buf) > pktsize: # True for any buffered data
data = buf[:pktsize] # empty (b'')
del buf[:pktsize] # no-op
...
self._send_window -= len(data) # -= 0, unchanged
With _send_pktsize == 0, pktsize is 0, so buf[:0] is empty,
del buf[:0] is a no-op, and _send_window is never decremented — the
while condition is permanently true, and with no await in the body the
event loop is blocked.
Impact
- Client vector (primary): a malicious SSH server replies to the client's
channel open with
maximum packet size = 0; the client wedges on its first channel write. The attacker is the server, so it needs no valid credentials. - Server vector: an authenticated client opens a channel with
maximum packet size = 0; any server-side channel write wedges the AsyncSSH server's event loop, freezing every current and future connection. A single low-privilege account can take the whole server down.
Both vectors are a single SSH message, deterministic, and cause total availability loss for the affected process.
Affected versions
<= 2.23.1 (latest release, 2026-06-06); also present on master
(channel.py:465/528 unguarded). Verified end-to-end on 2.23.1.
Verification
The maintainer's proposed fix (reject send_pktsize == 0 in connection.py
_process_channel_open / _process_channel_open_confirmation) was applied to
2.23.1 and re-tested end-to-end over TCP:
- Unpatched: malicious server (paramiko forcing
max_packet_size=0in OPEN_CONFIRMATION) + real asyncssh client → client event loop wedges. - Patched: the guard fires inside
_process_channel_open_confirmation, the malicious value is rejected, the connection closes cleanly (ChannelOpenError: SSH connection closed), and the client does not wedge.
The maintainer (Ron Frederick) independently confirmed the freeze and noted that even shutting the server down does not break clients out of the loop.
Reproducers available: a focused harness driving the real
SSHChannel._flush_send_buf with _send_pktsize=0, and an end-to-end
malicious_server.py (paramiko) + client.py (real asyncssh) pair. The
end-to-end client repro uses asyncio.new_event_loop() (not get_event_loop())
for Python 3.14 compatibility.
Suggested fix (maintainer's approach)
In connection.py, after each send_pktsize = packet.get_uint32() in
_process_channel_open and _process_channel_open_confirmation:
if send_pktsize == 0:
raise ProtocolError('Invalid maximum packet size')
CVSS
CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H (6.5 Medium). An earlier draft
quoted 7.5 High ("100% CPU"); the synchronous loop burns ~100% of one core's
worth of CPU but, being single-threaded, the OS scheduler spreads it across
cores, so the real impact is event-loop / connection freeze, not machine-wide
CPU exhaustion.
References
- RFC 4254 §5.1 (channel "maximum packet size"; behavior for 0 is undefined).
- The same
maximum packet size = 0send-loop wedge was confirmed in several other independent SSH implementations (different languages/runtimes) and reported to each maintainer separately.
Credits
Reported by zhangph (afldl), 2026-06-20. ```
{
"affected": [
{
"database_specific": {
"last_known_affected_version_range": "\u003c= 2.23.1"
},
"package": {
"ecosystem": "PyPI",
"name": "asyncssh"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "2.24.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-62949"
],
"database_specific": {
"cwe_ids": [
"CWE-835"
],
"github_reviewed": true,
"github_reviewed_at": "2026-09-17T14:53:01Z",
"nvd_published_at": "2026-09-16T20:17:26Z",
"severity": "MODERATE"
},
"details": "## Summary\n\nA malicious SSH server can wedge an AsyncSSH **client**, and an authenticated\nclient can wedge an AsyncSSH **server**, by sending a channel `maximum packet\nsize` of `0` in `SSH_MSG_CHANNEL_OPEN_CONFIRMATION` (server\u2192client) or\n`SSH_MSG_CHANNEL_OPEN` (client\u2192server). AsyncSSH stores the peer-supplied value\nverbatim with no lower-bound check; the first time channel data is written,\n`SSHChannel._flush_send_buf` enters a **synchronous infinite loop** that cannot\nbe interrupted by `asyncio.wait_for` or any timeout. The loop body has no\n`await`, so it blocks the entire asyncio event loop \u2014 for a server, one\nmalicious authenticated channel freezes **all** current and future connections.\n\nRFC 4254 \u00a75.1 leaves receiver behavior for a peer-reported \"maximum packet\nsize = 0\" undefined, so the value must be rejected rather than stored.\n\n## Root cause\n\n`asyncssh/channel.py`:\n\n```python\n# process_open (server side) -- line 465\nself._send_pktsize = send_pktsize # peer value, no \u003e= 1 check\n\n# process_open_confirmation (client) -- line 528\nself._send_pktsize = send_pktsize # peer value, no \u003e= 1 check\n\n# _flush_send_buf -- lines 305-320\nwhile self._send_buf and self._send_window:\n pktsize = min(self._send_window, self._send_pktsize) # 0 when peer sends 0\n buf, datatype = self._send_buf[0]\n if len(buf) \u003e pktsize: # True for any buffered data\n data = buf[:pktsize] # empty (b\u0027\u0027)\n del buf[:pktsize] # no-op\n ...\n self._send_window -= len(data) # -= 0, unchanged\n```\n\nWith `_send_pktsize == 0`, `pktsize` is `0`, so `buf[:0]` is empty,\n`del buf[:0]` is a no-op, and `_send_window` is never decremented \u2014 the\n`while` condition is permanently true, and with no `await` in the body the\nevent loop is blocked.\n\n## Impact\n\n- **Client vector (primary):** a malicious SSH server replies to the client\u0027s\n channel open with `maximum packet size = 0`; the client wedges on its first\n channel write. The attacker is the server, so it needs no valid credentials.\n- **Server vector:** an authenticated client opens a channel with\n `maximum packet size = 0`; any server-side channel write wedges the AsyncSSH\n server\u0027s event loop, freezing **every** current and future connection. A\n single low-privilege account can take the whole server down.\n\nBoth vectors are a single SSH message, deterministic, and cause total\navailability loss for the affected process.\n\n## Affected versions\n\n\u003c= 2.23.1 (latest release, 2026-06-06); also present on `master`\n(channel.py:465/528 unguarded). Verified end-to-end on 2.23.1.\n\n## Verification\n\nThe maintainer\u0027s proposed fix (reject `send_pktsize == 0` in `connection.py`\n`_process_channel_open` / `_process_channel_open_confirmation`) was applied to\n2.23.1 and re-tested end-to-end over TCP:\n\n- **Unpatched:** malicious server (paramiko forcing `max_packet_size=0` in\n OPEN_CONFIRMATION) + real asyncssh client \u2192 client event loop wedges.\n- **Patched:** the guard fires inside `_process_channel_open_confirmation`, the\n malicious value is rejected, the connection closes cleanly\n (`ChannelOpenError: SSH connection closed`), and the client does **not** wedge.\n\nThe maintainer (Ron Frederick) independently confirmed the freeze and noted that\n**even shutting the server down does not break clients out of the loop**.\n\nReproducers available: a focused harness driving the real\n`SSHChannel._flush_send_buf` with `_send_pktsize=0`, and an end-to-end\n`malicious_server.py` (paramiko) + `client.py` (real asyncssh) pair. The\nend-to-end client repro uses `asyncio.new_event_loop()` (not `get_event_loop()`)\nfor Python 3.14 compatibility.\n\n## Suggested fix (maintainer\u0027s approach)\n\nIn `connection.py`, after each `send_pktsize = packet.get_uint32()` in\n`_process_channel_open` and `_process_channel_open_confirmation`:\n\n```python\nif send_pktsize == 0:\n raise ProtocolError(\u0027Invalid maximum packet size\u0027)\n```\n\n## CVSS\n\n`CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H` (6.5 Medium). An earlier draft\nquoted 7.5 High (\"100% CPU\"); the synchronous loop burns ~100% of one core\u0027s\nworth of CPU but, being single-threaded, the OS scheduler spreads it across\ncores, so the real impact is event-loop / connection freeze, not machine-wide\nCPU exhaustion.\n\n## References\n\n- RFC 4254 \u00a75.1 (channel \"maximum packet size\"; behavior for 0 is undefined).\n- The same `maximum packet size = 0` send-loop wedge was confirmed in several\n other independent SSH implementations (different languages/runtimes) and\n reported to each maintainer separately.\n\n## Credits\n\nReported by zhangph (afldl), 2026-06-20.\n```",
"id": "GHSA-rw4j-r22c-9gc3",
"modified": "2026-09-17T14:53:01Z",
"published": "2026-09-17T14:53:01Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/ronf/asyncssh/security/advisories/GHSA-rw4j-r22c-9gc3"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-62949"
},
{
"type": "WEB",
"url": "https://github.com/ronf/asyncssh/commit/756cbae5350789ce9735f15f704bae9b5a3608b8"
},
{
"type": "WEB",
"url": "https://github.com/ronf/asyncssh/commit/9c354270c009285525e126721e8ed5fbed1f8a67"
},
{
"type": "PACKAGE",
"url": "https://github.com/ronf/asyncssh"
},
{
"type": "WEB",
"url": "https://github.com/ronf/asyncssh/releases/tag/v2.24.0"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
],
"summary": "AsyncSSH: asyncio event-loop freeze via SSH maximum packet size = 0 in SSH_MSG_CHANNEL_OPEN / OPEN_CONFIRMATION"
}
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.