xion-oauth2
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill scripts (login.sh, logout.sh, refresh.sh, and status.sh) interact with the system via the
xion-toolkitCLI. Analysis confirms that all command arguments are correctly handled using Bash arrays and quoted expansions (e.g.,"${CMD_ARGS[@]}"), which prevents shell injection vulnerabilities even when processing user-supplied parameters like network names or port numbers. - [EXTERNAL_DOWNLOADS]: The skill requires the
xion-toolkitbinary to function. It properly manages this dependency by requiring thexion-toolkit-initskill for installation, ensuring that the CLI is sourced from the vendor's official distribution channel. - [SAFE]: The skill is designed for credential management but does not handle plaintext secrets within its own logic. As documented and verified in the scripts, authentication tokens are managed by the underlying CLI tool and stored securely in the host's native keyring, minimizing the risk of exposure in the agent's environment.
Audit Metadata