{"id":"CVE-2026-50144","aliases":["EUVD-2026-44794","GHSA-jxmc-3mv6-7pwr"],"status":"published","title":"ncnn: Out-of-bounds heap write in ParamDict::load_param via unchecked negative parameter id","description":"ncnn is a high-performance neural network inference framework optimized for the mobile platform. In commit e54f7b1f88434e1d844ea0551b880a1cfb079ce1 and earlier, ncnn allows an out-of-bounds heap write in ncnn::ParamDict::load_param() when Net::load_param() loads a malicious .param model file because the parsed parameter id is checked only against id \u003e= NCNN_MAX_PARAM_COUNT, allowing a negative id to index before the params[NCNN_MAX_PARAM_COUNT] array. This vulnerability is fixed by commit 5a0288f255daa6c3294f77109f67718e434ec020.","severity":{"version":"3.1","score":7.1,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H","level":"high","assessor":"cna","from":"cve"},"epss":{"day":"2026-11-09","score":0.00161,"percentile":0.05305,"model":"v2026.06.15"},"facts":{"in_kev":false,"known_exploited":false,"public_exploit":false,"fix_available":false},"cwes":["CWE-129","CWE-20","CWE-787"],"published":"2026-07-15T20:04:07Z","modified":"2026-10-28T10:51:55Z","products":[{"vendor":"tencent","product":"ncnn"}],"references":[{"url":"https://github.com/Tencent/ncnn/security/advisories/GHSA-jxmc-3mv6-7pwr","tags":["x_refsource_CONFIRM","exploit","advisory"]},{"url":"https://github.com/Tencent/ncnn/commit/5a0288f255daa6c3294f77109f67718e434ec020","tags":["x_refsource_MISC","fix"]},{"url":"https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/50xxx/CVE-2026-50144.json","tags":["advisory"]},{"url":"https://nvd.nist.gov/vuln/detail/CVE-2026-50144","tags":["advisory"]}],"sources":{"cve":[{"id":"CVE-2026-50144","title":"ncnn: Out-of-bounds heap write in ParamDict::load_param via unchecked negative parameter id","description":"ncnn is a high-performance neural network inference framework optimized for the mobile platform. In commit e54f7b1f88434e1d844ea0551b880a1cfb079ce1 and earlier, ncnn allows an out-of-bounds heap write in ncnn::ParamDict::load_param() when Net::load_param() loads a malicious .param model file because the parsed parameter id is checked only against id \u003e= NCNN_MAX_PARAM_COUNT, allowing a negative id to index before the params[NCNN_MAX_PARAM_COUNT] array. This vulnerability is fixed by commit 5a0288f255daa6c3294f77109f67718e434ec020.","status":"published","published":"2026-07-15T20:04:07.099Z","modified":"2026-07-17T12:27:10.432Z","cvss":[{"version":"3.1","score":7.1,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H","level":"high","assessor":"cna"}],"cwes":["CWE-20","CWE-129","CWE-787"],"updated_at":"2026-11-02T14:20:09.098623Z"}],"euvd":[{"id":"EUVD-2026-44794","description":"ncnn is a high-performance neural network inference framework optimized for the mobile platform. In commit e54f7b1f88434e1d844ea0551b880a1cfb079ce1 and earlier, ncnn allows an out-of-bounds heap write in ncnn::ParamDict::load_param() when Net::load_param() loads a malicious .param model file because the parsed parameter id is checked only against id \u003e= NCNN_MAX_PARAM_COUNT, allowing a negative id to index before the params[NCNN_MAX_PARAM_COUNT] array. This vulnerability is fixed by commit 5a0288f255daa6c3294f77109f67718e434ec020.","status":"published","published":"2026-07-15T20:04:07Z","modified":"2026-07-17T12:27:10Z","cvss":[{"version":"3.1","score":7.1,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H","level":"high"}],"updated_at":"2026-11-02T18:19:50.846094Z"}],"nvd":[{"id":"CVE-2026-50144","description":"ncnn is a high-performance neural network inference framework optimized for the mobile platform. In commit e54f7b1f88434e1d844ea0551b880a1cfb079ce1 and earlier, ncnn allows an out-of-bounds heap write in ncnn::ParamDict::load_param() when Net::load_param() loads a malicious .param model file because the parsed parameter id is checked only against id \u003e= NCNN_MAX_PARAM_COUNT, allowing a negative id to index before the params[NCNN_MAX_PARAM_COUNT] array. This vulnerability is fixed by commit 5a0288f255daa6c3294f77109f67718e434ec020.","status":"published","published":"2026-07-15T20:17:13.5Z","modified":"2026-10-24T21:10:00.13Z","cvss":[{"version":"3.1","score":7.1,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H","level":"high","assessor":"secondary"}],"cwes":["CWE-20","CWE-129","CWE-787"],"updated_at":"2026-11-02T14:12:29.879716Z"}],"osv":[{"id":"CVE-2026-50144","title":"ncnn: Out-of-bounds heap write in ParamDict::load_param via unchecked negative parameter id","description":"ncnn is a high-performance neural network inference framework optimized for the mobile platform. In commit e54f7b1f88434e1d844ea0551b880a1cfb079ce1 and earlier, ncnn allows an out-of-bounds heap write in ncnn::ParamDict::load_param() when Net::load_param() loads a malicious .param model file because the parsed parameter id is checked only against id \u003e= NCNN_MAX_PARAM_COUNT, allowing a negative id to index before the params[NCNN_MAX_PARAM_COUNT] array. This vulnerability is fixed by commit 5a0288f255daa6c3294f77109f67718e434ec020.","status":"published","published":"2026-07-15T20:04:07.099Z","modified":"2026-10-28T10:51:55.977308611Z","cvss":[{"version":"3.1","score":7.1,"vector":"CVSS:3.1/AV:L/AC:L/PR:N/UI:R/S:U/C:N/I:H/A:H","level":"high"}],"cwes":["CWE-129","CWE-20","CWE-787"],"updated_at":"2026-11-02T15:45:07.060148Z"}]}}