{"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/cpe2.3acozecoze_studio/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:coze:coze_studio:*:*:*:*:*:*:*:*"],"_cs_cves":[{"cvss":8.8,"id":"CVE-2026-77934"}],"_cs_exploited":false,"_cs_has_poc":false,"_cs_poc_references":[],"_cs_products":["Coze Studio (\u003c 0.5.0)"],"_cs_severities":["high"],"_cs_tags":["remote-code-execution","sandbox-escape","python"],"_cs_type":"advisory","_cs_vendors":["Coze"],"content_html":"\u003cp\u003eCoze Studio versions prior to 0.5.0 contain an insecure default configuration that fails to properly sandbox Python code execution within workflow nodes. An authenticated attacker can manipulate these workflow nodes to execute arbitrary code within a subprocess. By bypassing the existing regex-based import filters - specifically using the \u003ccode\u003e__import__('subprocess')\u003c/code\u003e technique - an attacker can gain command execution on the host environment. This vulnerability poses a significant risk to the integrity and confidentiality of the Coze Studio instance, as successful exploitation allows an attacker to exfiltrate sensitive environment variables, extract model API keys, and pivot to connected backend services such as MySQL, MinIO, Redis, and etcd.\u003c/p\u003e\n\u003ch2 id=\"impact\"\u003eImpact\u003c/h2\u003e\n\u003cp\u003eSuccessful exploitation of CVE-2026-77934 enables an attacker to achieve remote code execution as the user running the Coze Studio service. This access facilitates the theft of credentials (API keys) and unauthorized access to internal data stores (MySQL, MinIO, Redis, etcd), potentially resulting in full database compromise or cross-environment lateral movement.\u003c/p\u003e\n\u003ch2 id=\"recommendation\"\u003eRecommendation\u003c/h2\u003e\n\u003col\u003e\n\u003cli\u003eUpgrade Coze Studio to version 0.5.0 or later to ensure proper sandboxing of Python code execution.\u003c/li\u003e\n\u003cli\u003eReview logs for unexpected \u003ccode\u003esubprocess\u003c/code\u003e or \u003ccode\u003eos.system\u003c/code\u003e calls originating from the service process.\u003c/li\u003e\n\u003cli\u003eRestrict network access for the Coze Studio backend infrastructure to prevent unauthorized access to internal databases (MySQL, Redis, etcd, MinIO) if the process is compromised.\u003c/li\u003e\n\u003c/ol\u003e\n","date_modified":"2026-10-20T02:32:17Z","date_published":"2026-10-20T02:32:17Z","id":"https://feed.craftedsignal.io/briefs/2026-10-coze-studio-rce/","summary":"Coze Studio versions prior to 0.5.0 allow authenticated users to execute arbitrary commands by bypassing sandbox protections in Python workflow nodes.","title":"Remote Code Execution in Coze Studio via Workflow Nodes","url":"https://feed.craftedsignal.io/briefs/2026-10-coze-studio-rce/"}],"language":"en","title":"CraftedSignal Threat Feed - Cpe:2.3:a:coze:coze_studio:*:*:*:*:*:*:*:*","version":"https://jsonfeed.org/version/1.1"}