https://cve.radiocsirt.org Contains only the most 10 recent entries. http://www.rssboard.org/rss-specification python-feedgen en Fri, 02 Oct 2026 13:41:07 +0000 https://cve.radiocsirt.org/vuln/euvd-2026-375149 EUVD-2026-375149 https://cve.radiocsirt.org/vuln/euvd-2026-375149 https://cve.radiocsirt.org/vuln/fkie_cve-2026-93840 <p>vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.</p> <p>vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.</p> https://cve.radiocsirt.org/vuln/fkie_cve-2026-93840 https://cve.radiocsirt.org/vuln/ghsa-97fp-7cr2-23qc <p>vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.</p> <p>vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.</p> https://cve.radiocsirt.org/vuln/ghsa-97fp-7cr2-23qc https://cve.radiocsirt.org/vuln/pysec-2026-3998 <p><strong>Affected:</strong> PyPI: vllm</p> <p>vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.</p> <p><strong>Affected:</strong> PyPI: vllm</p> <p>vLLM before 0.29.0 validates allowed_token_ids against tokenizer length instead of model output logits width in SamplingParams._validate_allowed_token_ids(). Attackers can supply token IDs above the output vocabulary that pass validation, causing LogitBiasState to corrupt GPU logits state and allow concurrent requests to sample tokens outside their allowlists.</p> https://cve.radiocsirt.org/vuln/pysec-2026-3998 https://cve.radiocsirt.org/vuln/wid-sec-w-2026-3483 <p>Ein Angreifer kann mehrere Schwachstellen in vllm ausnutzen, um Daten zu manipulieren.</p> <p>Ein Angreifer kann mehrere Schwachstellen in vllm ausnutzen, um Daten zu manipulieren.</p> https://cve.radiocsirt.org/vuln/wid-sec-w-2026-3483