oceanbase-deploy

Warn

Audited by Gen Agent Trust Hub on Sep 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONPERSISTENCEDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill extensively uses shell commands via the 'obd' (OceanBase Deployer) binary to perform database operations, cluster lifecycle management, and host initialization.- [DYNAMIC_EXECUTION]: The file 'cluster-management/scripts/obd-edit-config-pty.py' is a utility script that uses 'pty.fork()' and 'os.execve()' to execute the 'obd' binary. While the script incorporates security controls such as SHA-256 integrity verification of the binary and environment variable scrubbing (removing LD_PRELOAD, etc.), wrapping executables in this manner is a significant capability.- [PRIVILEGE_ESCALATION]: Skill instructions guide the agent to use 'sudo' for high-privilege tasks, including RPM package installation, modifying system limits (nofile, nproc), and configuring kernel parameters (vm.max_map_count) through 'obd host init' or manual remediation.- [PERSISTENCE]: The skill automates the setup of passwordless SSH access from the controller to managed nodes by appending the controller's public key to the '~/.ssh/authorized_keys' file on target hosts.- [EXTERNAL_DOWNLOADS]: The skill downloads the 'obd' controller (ob-deploy) and database components from 'mirrors.oceanbase.com' and 'mirrors.aliyun.com'. These are official vendor mirrors and well-known services, which are documented neutrally in accordance with the vendor context.- [INDIRECT_PROMPT_INJECTION]: The skill has a large attack surface for indirect injection as it ingests and processes untrusted external data, including user-supplied YAML configuration files, benchmark scripts (.test files), and remote server logs, which are then used to influence agent actions and command generation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 16, 2026, 06:25 AM