[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"repo-stars":3,"vuln-CVE-2026-55743":6},{"stargazers_count":4,"fetched_at":5},7,"2026-06-21T07:39:11.499Z",{"id":7,"descriptions":8,"cisa":9,"weaknesses":10,"exploits":76,"aliases":77,"duplicate_of":9,"upstream":78,"downstream":79,"duplicates":80,"related":81,"reserved_at":9,"published_at":82,"modified_at":83,"state":84,"summary":85,"references_raw":94,"kevs":110,"epss":111,"epss_history":114,"metrics":122,"affected":131},"CVE-2026-55743","The shell tool command allowlist in the SecurityPolicy of OpenHuman desktop agent through 0.54.0 (default Supervised security policy) can be bypassed to execute arbitrary OS commands with the privileges of the desktop user. Two flaws in src/openhuman/security/policy.rs combine: (1) is_args_safe() blocks the find flags -exec and -ok but not the functionally identical -execdir and -okdir, which also execute an arbitrary command for each matched file; and (2) skip_env_assignments() strips leading inline KEY=value environment-variable assignments before allowlist validation, so a command such as GIT_EXTERNAL_DIFF=\u003Ccmd> git diff is validated as the allowed git diff but, when executed via the shell, runs \u003Ccmd> through git's environment-driven hooks (for example GIT_EXTERNAL_DIFF or GIT_SSH_COMMAND). Because the sandbox is the primary trust boundary between untrusted LLM-processed content and the host operating system, an attacker can achieve remote code execution via indirect prompt injection: a malicious document, email, calendar event, or web page ingested by the agent instructs it to run a benign-looking allowlisted command, resulting in arbitrary command execution, data exfiltration, arbitrary file read/write, and lateral movement on the user's machine. The issue was fixed in commit 60050aa09a870f53ed7e4cd40ed41fd2860329e7 (first released in 0.54.22-staging; first stable release 0.56.0), which blocks -execdir/-okdir for find.",null,[11,40],{"_key":12,"id":12,"name":13,"description":14,"type":15,"status":16,"abstraction":17,"likelihood_of_exploit":18,"capec":19},"CWE-78","Improper Neutralization of Special Elements used in an OS Command ('OS Command Injection')","The product constructs all or part of an OS command using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the intended OS command when it is sent to a downstream component.","weakness","Stable","Base","High",[20,24,28,32,36],{"id":21,"name":22,"techniques":23},"CAPEC-108","Command Line Execution through SQL Injection",[],{"id":25,"name":26,"techniques":27},"CAPEC-15","Command Delimiters",[],{"id":29,"name":30,"techniques":31},"CAPEC-43","Exploiting Multiple Input Interpretation Layers",[],{"id":33,"name":34,"techniques":35},"CAPEC-6","Argument Injection",[],{"id":37,"name":38,"techniques":39},"CAPEC-88","OS Command Injection",[],{"_key":41,"id":41,"name":42,"description":43,"type":15,"status":44,"abstraction":17,"likelihood_of_exploit":9,"capec":45},"CWE-184","Incomplete List of Disallowed Inputs","The product implements a protection mechanism that relies on a list of inputs (or properties of inputs) that are not allowed by policy or otherwise require other action to neutralize before additional processing takes place, but the list is incomplete.","Draft",[46,50,52,56,60,62,64,68,72],{"id":47,"name":48,"techniques":49},"CAPEC-120","Double Encoding",[],{"id":25,"name":26,"techniques":51},[],{"id":53,"name":54,"techniques":55},"CAPEC-182","Flash Injection",[],{"id":57,"name":58,"techniques":59},"CAPEC-3","Using Leading 'Ghost' Character Sequences to Bypass Input Filters",[],{"id":29,"name":30,"techniques":61},[],{"id":33,"name":34,"techniques":63},[],{"id":65,"name":66,"techniques":67},"CAPEC-71","Using Unicode Encoding to Bypass Validation Logic",[],{"id":69,"name":70,"techniques":71},"CAPEC-73","User-Controlled Filename",[],{"id":73,"name":74,"techniques":75},"CAPEC-85","AJAX Footprinting",[],[],[],[],[],[],[],"2026-06-17T14:08:33.726Z","2026-06-17T15:40:47.796Z","PUBLISHED",{"cisa_kev":86,"cisa_ransomware":86,"cisa_vendor":9,"epss_severity":87,"epss_score":88,"severity":89,"severity_score":90,"severity_version":91,"severity_source":92,"severity_vector":93,"severity_status":84},false,"low",0.00704,"critical",9.6,"v3.1","cve.org","CVSS:3.1/AV:N/AC:L/PR:N/UI:R/S:C/C:H/I:H/A:H",[95,100,105],{"url":96,"sources":97,"tags":98},"https://github.com/tinyhumansai/openhuman/commit/60050aa09a870f53ed7e4cd40ed41fd2860329e7",[92],[99],"Patch",{"url":101,"sources":102,"tags":103},"https://github.com/tinyhumansai/openhuman/blob/v0.53.49-staging/src/openhuman/security/policy.rs",[92],[104],"Technical Description",{"url":106,"sources":107,"tags":108},"https://github.com/tinyhumansai/openhuman",[92],[109],"Product",[],{"date":112,"score":88,"percentile":113},"2026-06-20",0.48421,[115,118,121],{"date":116,"score":88,"percentile":117},"2026-06-18",0.48429,{"date":119,"score":88,"percentile":120},"2026-06-19",0.48424,{"date":112,"score":88,"percentile":113},[123],{"source":92,"cvss_v2_0":9,"cvss_v3_0":9,"cvss_v3_1":124,"cvss_v4_0":128},{"baseScore":90,"baseSeverity":125,"vectorString":93,"impactScore":126,"exploitabilityScore":127},"CRITICAL",10,7.2,{"baseScore":129,"baseSeverity":125,"vectorString":130,"impactScore":9,"exploitabilityScore":9},9.4,"CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H",[132],{"ecosystem":9,"name":133,"vendor":134,"product":135,"cpe_part":136,"purl_type":9,"purl_namespace":9,"purl_name":9,"source":9,"versions":137},"OpenHuman","tinyhumansai","openhuman","a",[138],{"version":139,"is_range":140,"range_type":92,"version_start":9,"version_start_type":9,"version_end":141,"version_end_type":142,"fixed_in":9},"\u003C= 0.54.0",true,"0.54.0","including"]