home-assistant
Pass
Audited by Gen Agent Trust Hub on Apr 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local shell scripts (status.sh, query.sh, control.sh) and the mcporter CLI to interact with the Home Assistant REST API and MCP server. It includes an option to bypass SSL verification (using curl -k) for connections to local instances.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its ingestion of external data from the Home Assistant API.
- Ingestion points: Entity names, states, attributes, and configuration metadata retrieved from the Home Assistant API or MCP tools.
- Boundary markers: Absent; the instructions do not define delimiters or provide warnings to the agent regarding embedded instructions in the retrieved data.
- Capability inventory: The skill allows for comprehensive control of smart home devices, including the ability to trigger automations, execute scripts, and modify system configurations (e.g., dashboards and automations).
- Sanitization: Absent; data retrieved from the external Home Assistant instance is processed directly without specified validation or filtering.
Audit Metadata