PraisonAI before 4.5.128 contains an arbitrary shell command execution vulnerability where the UI modules hardcode approvalmode to auto, overriding administrator configuration from PRAISONAPPROVAL_MODE environment variable. Authenticated attackers can instruct the LLM agent to execute arbitrary shell commands via subprocess.run with shell=True, bypassing the manual approval gate and insufficient command sanitization blocklists.
{
"osv_generated_from": "https://github.com/CVEProject/cvelistV5/tree/main/cves/2026/56xxx/CVE-2026-56075.json",
"cwe_ids": [
"CWE-863"
],
"cna_assigner": "VulnCheck"
}