{"description":"Trending threats, MITRE ATT\u0026CK coverage, and detection metadata. Fed continuously.","favicon":"https://feed.craftedsignal.io/favicon-32x32.png","feed_url":"https://feed.craftedsignal.io/cpes/cpe2.3allama.cppllama.cpp/feed.json","home_page_url":"https://feed.craftedsignal.io/","icon":"https://feed.craftedsignal.io/apple-touch-icon.png","items":[{"_cs_actors":[],"_cs_cpes":["cpe:2.3:a:llama.cpp:llama.cpp:*:*:*:*:*:*:*:*"],"_cs_cves":[{"cvss":7.8,"id":"CVE-2026-116816"}],"_cs_exploited":false,"_cs_has_poc":false,"_cs_poc_references":[],"_cs_products":["llama.cpp (b5899 \u003c= version \u003c b10261)"],"_cs_severities":["high"],"_cs_tags":["heap-overflow","vulnerability","code-execution"],"_cs_type":"advisory","_cs_vendors":["llama.cpp"],"content_html":"\u003cp\u003eThe llama.cpp library contains a heap out-of-bounds write vulnerability within the \u003ccode\u003ellm_tokenizer_plamo2::build\u003c/code\u003e function. An attacker can supply a crafted GGUF model file containing BYTE-typed token text with negative parsed values (e.g., \u003ccode\u003e\u0026lt;0x-1\u0026gt;\u003c/code\u003e). When the model is loaded, the application attempts to write token IDs into memory before the intended \u003ccode\u003ebytes_\u003c/code\u003e buffer, resulting in heap corruption.\u003c/p\u003e\n\u003ch2 id=\"impact\"\u003eImpact\u003c/h2\u003e\n\u003cp\u003eSuccessful exploitation may lead to application crashes or arbitrary code execution within the context of the llama.cpp process. The impact depends on the environment in which the model is being loaded.\u003c/p\u003e\n\u003ch2 id=\"recommendation\"\u003eRecommendation\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eUpdate llama.cpp to version b10261 or later immediately to incorporate the patch for CVE-2026-116816.\u003c/li\u003e\n\u003cli\u003eImplement strict input validation or sandboxing for GGUF model files loaded from untrusted sources.\u003c/li\u003e\n\u003cli\u003eAudit environments that ingest arbitrary GGUF models for unauthorized or unexpected process crashes, which may indicate exploitation attempts.\u003c/li\u003e\n\u003c/ul\u003e\n","date_modified":"2026-10-27T01:52:36Z","date_published":"2026-10-27T01:52:36Z","id":"https://feed.craftedsignal.io/briefs/2026-10-llama-cpp-heap-overflow/","summary":"A heap out-of-bounds write vulnerability in llama.cpp versions b5899 through b10260 allows attackers to achieve memory corruption or potential code execution via maliciously crafted GGUF model files.","title":"llama.cpp: Heap out-of-bounds write in llm_tokenizer_plamo2 (CVE-2026-116816)","url":"https://feed.craftedsignal.io/briefs/2026-10-llama-cpp-heap-overflow/"}],"language":"en","title":"CraftedSignal Threat Feed - Cpe:2.3:a:llama.cpp:llama.cpp:*:*:*:*:*:*:*:*","version":"https://jsonfeed.org/version/1.1"}