ha-dev
Pass
Audited by Gen Agent Trust Hub on Jun 29, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides educational content and boilerplate code for Home Assistant development. All examples involving sensitive data, such as API tokens or internal URLs, use clear placeholders (e.g., , , YOUR_API_KEY_HERE) or environment variables, adhering to safe coding practices.\n- [EXTERNAL_DOWNLOADS]: Mentions well-known Home Assistant community resources and tools, including HACS integrations (Mushroom Cards, Bubble Card) and the mcp-proxy tool. These are recognized services within the ecosystem and the skill provides instructions for manual setup by the user.\n- [COMMAND_EXECUTION]: Provides documentation on using standard command-line tools such as curl for API interactions, grep for searching configuration files, and ha core for system management. These commands are intended for manual execution by the developer for debugging and configuration purposes.\n- [REMOTE_CODE_EXECUTION]: References the mcp-proxy tool via its official GitHub repository for environment setup. This is a standard developer practice and does not involve the skill executing untrusted remote code autonomously.\n- [PROMPT_INJECTION]: The skill documents how Home Assistant processes data from various sources (sensors, webhooks, and external APIs). It mitigates potential risks by recommending best practices such as guarding templates against invalid states (unavailable, unknown) and using strict schema validation (voluptuous) for custom services.\n
- Ingestion points: Sensor state triggers, webhook payloads, and external API responses via integration coordinators (reference/automations.md, reference/integrations.md).\n
- Boundary markers: Recommends template guards for entity states and explicit variable definitions in blueprints (SKILL.md, reference/automations.md).\n
- Capability inventory: Service calls (acting on devices), shell command execution for maintenance, and network requests via httpx for integrations.\n
- Sanitization: Encourages type casting in Jinja2 templates and validation schemas for service calls.
Audit Metadata