CVE-2026-76850: InternLM lmdeploy · Deserialization of Untrusted Data
Affects InternLM lmdeploySource CVE Database V5
Not known to be exploited · CVSS 9.3 Critical : upgrade to 0.16.0 as a priority.
Data as of 3 Nov 2026, 22:58 UTC
- EPSS
- 1.2% chance of exploitation in 30 days – higher than 69% of CVEsExploit Prediction Scoring System: the likelihood attackers will use this flaw soon.
- Patch
- Available – fixed in 0.16.0
CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:NExploitation evidence
Disclosed: evidence found.
CVE published 19 Aug 2026
Detectable: no evidence on record.
No public nuclei template indexed
Weaponized: no evidence on record.
No public exploit code indexed in Metasploit or Exploit-DB
Exploited in the wildExploited: no evidence on record.
Not in CISA KEV and no source reports exploitation
Observed by OffSeq MirageSeen by Mirage: no evidence on record.
No attack traffic recorded by OffSeq Mirage sensors
A missing step is not evidence of safety. Sources: cve.org (opens in a new tab), the Metasploit, Exploit-DB and nuclei-templates index as of 3 Nov 2026.
What to do
Upgrade InternLM lmdeploy to 0.16.0.
Affected software
| Product | Affected | Fixed in |
|---|---|---|
| InternLMlmdeploy | Affected
| Fixed in
|
PyPIlmdeploypkg:pypi/lmdeployOther threats | Affected
| Fixed in
|
Check your own hosts for these packages with threat-finder (opens in a new tab), our free open-source scanner.
AI-suggested
Users should enable api_keys authentication to restrict access to the ZeroMQ endpoints or disable disaggregated serving to avoid exposure.
Description
LMDeploy deserializes disaggregated-serving peer messages with pickle. The handle_zmq_recv coroutine in lmdeploy/pytorch/disagg/conn/engine_conn.py reads peer-to-peer cache-free requests with recv_pyobj(), which deserializes the received bytes with pickle.loads(), and the isinstance check against DistServeCacheFreeRequest runs only after deserialization has already completed. The peer that supplies those bytes is caller-controlled: p2p_connect passes remote_engine_endpoint_info.zmq_address from the request body to connect() on the ZMQ PULL socket, and the POST /distserve/p2p_initialize and /distserve/p2p_connect endpoints in lmdeploy/serve/openai/api_server.py apply no authentication unless the server is started with api_keys, which defaults to None. A remote attacker can direct an engine to pull from a ZMQ endpoint under their control and execute arbitrary code in the engine process. Deployments that do not enable disaggregated serving are not affected, because the receive loop is only started once the migration backend accepts the connection.
References
- github.com/InternLM/lmdeploy/issues/4804 (opens in a new tab)
- github.com/InternLM/lmdeploy/commit/f05b4ad8bf2e2d84101a1d63b3c44fadd99223b2 (opens in a new tab)
- github.com/InternLM/lmdeploy/blob/v0.15.0/lmdeploy/pytorch/disagg/conn/engine_conn.py (opens in a new tab)
- github.com/InternLM/lmdeploy/blob/v0.15.0/lmdeploy/pytorch/disagg/conn/engine_conn.py (opens in a new tab)
- github.com/InternLM/lmdeploy/releases/tag/v0.16.0 (opens in a new tab)
- github.com/InternLM/lmdeploy (opens in a new tab)
- vulncheck.com/advisories/lmdeploy-remote-code-execution-via-unsafe-pickle-deserialization-in-the-disaggregated-serving-peer-connector (opens in a new tab)