haos-https-tunnel
Warn
Audited by Gen Agent Trust Hub on Oct 5, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill adds a third-party add-on repository from a personal GitHub account (https://github.com/brenner-tobias/ha-addons) to install the cloudflared add-on.
- [REMOTE_CODE_EXECUTION]: The skill installs and executes code from an untrusted remote source by adding an external repository to the Home Assistant supervisor.
- [COMMAND_EXECUTION]: The skill executes multiple administrative shell commands via SSH, including repository management and system API calls via curl.
- [PRIVILEGE_ESCALATION]: The instructions direct the user to disable Protection mode for the SSH add-on and use sudo for file system modifications, granting the agent elevated control over the host operating system.
- [INDIRECT_PROMPT_INJECTION]: The skill processes log output from the cloudflared add-on, which contains data from external services. Ingestion points: Application logs (SKILL.md). Boundary markers: Absent. Capability inventory: SSH command execution, privileged file writing, and Supervisor API access. Sanitization: Absent.
Audit Metadata