home-assistant-automation-scripts

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill instructions enable the agent to ingest external, potentially untrusted data from a Home Assistant instance and use it to perform actions with side effects.
  • Ingestion points: The skill uses tools like ha_get_overview, ha_search_entities, ha_config_get_automation, ha_get_logbook, ha_get_history, and ha_get_statistics to pull data into the context.
  • Boundary markers: Absent. There are no delimiters or specific instructions for the agent to disregard instructions that might be embedded in entity names, attributes, or log entries.
  • Capability inventory: The skill can create or update configurations via ha_config_set_automation and ha_config_set_script, import logic via ha_import_blueprint, and execute dynamic logic via ha_eval_template.
  • Sanitization: Absent. No validation or sanitization rules are defined to check the integrity or safety of the retrieved data before it influences the agent's logic or configuration changes.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:00 PM