The Agent Skills Directory

[SAFE] (SAFE): The skill's content is entirely focused on providing Home Assistant configuration examples and documentation. No obfuscation, data exfiltration patterns, or persistence mechanisms were detected.
[COMMAND_EXECUTION] (LOW): The SKILL.md file contains benign troubleshooting commands intended for the user, such as ha core logs and a Python one-liner using yaml.safe_load() to validate configuration files. These are standard diagnostic practices for the target platform.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill documents the use of Markdown cards with Jinja2 templates. This represents a standard vulnerability surface where untrusted entity data could theoretically influence the rendering environment.
Ingestion points: Markdown card content field in references/card-reference.md and assets/card-snippets.yaml.
Boundary markers: None (Standard HA template syntax used).
Capability inventory: Write, Edit, and WebFetch tools are authorized in the skill metadata.
Sanitization: Not explicitly defined within the skill; relies on Home Assistant's internal security model.

ha-dashboard