Xinference vulnerable to remote code execution via unsafe `eval()` in Llama3 tool-call parsing

SummaryXinference used Python's unsafe eval function when parsing Llama3 tool-call output generated by a large language model. Because the model output can be influenced by attacker-controlled prompts sent to the chat completion API, a remote attacker can craft prompts that cause the model to...

PYSEC-2026-3946 Xinference vulnerable to remote code execution via unsafe `eval()` in Llama3 tool-call parsing

Summary Xinference used Python's unsafe eval function when parsing Llama3 tool-call output generated by a large language model. Because the model output can be influenced by attacker-controlled prompts sent to the chat completion API, a remote attacker can craft prompts that cause the model to...

CVE-2026-85668

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,...

CVE-2026-85668 Xinference 3.3.0 Unauthenticated Arbitrary-Path File Read via /v1/models/llm/auto-register

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,...

CVE-2026-85668 Xinference 3.3.0 Unauthenticated Arbitrary-Path File Read via /v1/models/llm/auto-register

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,...

CVE-2026-85668

Xinference (v3.x, affected commit 4a94832, specifically v3.3.0) contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint. The endpoint accepts a caller-supplied model_path parameter without authentication or path confinement, reads and p...

CVE-2026-85668 Xinference 3.3.0 Unauthenticated Arbitrary-Path File Read via /v1/models/llm/auto-register

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,...

EUVD-2026-71124

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,...

PT-2026-85561

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied model path parameter without authentication or path confinement. The endpoint reads and parses config.json...

CVE-2026-85668: External Control of File Name or Path

Xinference affected commit 4a94832, v3.x contains an unauthenticated arbitrary-path file read vulnerability in the POST /v1/models/llm/auto-register endpoint, which accepts a caller-supplied modelpath parameter without authentication or path confinement. The endpoint reads and parses config.json,...

CVE-2026-76841

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

CVE-2026-76841

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

CVE-2026-76841 Xinference through 2.11.0 Remote Code Execution via Hardcoded trust_remote_code in Model Loaders

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

CVE-2026-76841 Xinference through 2.11.0 Remote Code Execution via Hardcoded trust_remote_code in Model Loaders

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

CVE-2026-76841

Xinference (by XorbitsAI) through version 2.11.0 unconditionally sets trust_remote_code=True at six model loader call sites spanning the rerank, embedding, and LLM modules. An attacker with model launch access can register a model with an unknown type and an arbitrary path, causing AutoTokenizer....

EUVD-2026-64847

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

CVE-2026-76841 Xinference through 2.11.0 Remote Code Execution via Hardcoded trust_remote_code in Model Loaders

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

PT-2026-80851

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trust remote code=True as a literal or as an unconditional default: RerankModel. get tokenizer in...

CVE-2026-76841: Improper Control of Generation of Code

Xinference loads models with Hugging Face remote code execution unconditionally enabled, and before version 2.12.0 exposes no setting to disable it. Six loader call sites pass trustremotecode=True as a literal or as an unconditional default: RerankModel.gettokenizer in...

CVE-2026-61539

Xinference is an inference API for running open-source, speech, and multimodal models. In 2.5.0 and earlier, Xinference passes attacker-influenced Llama3 tool-call output to eval in xinference/model/llm/toolparsers/llama3toolparser.py and xinference/model/llm/utils.py. Requests to...